ML14114A550

Joseph M. Farley Nuclear Plant, Response to Request for Additional Information Regarding License Amendment Request for Transition to 10 CFR 50.48(c) - NFPA 805 Performance Based Standard for Fire Protection for Light Water Reactor Generatin

ML14114A550
Person / Time
Site:	Farley
Issue date:	04/23/2014
From:	Pierce C R Southern Co, Southern Nuclear Operating Co
To:	Document Control Desk, Office of Nuclear Reactor Regulation
References
NL-14-0446
Download: ML14114A550 (63)

Text

Charles R. Pierce Southern NuclearRegulatory Affairs Director Operating

Company, Inc.40 Inverness Center ParkwayPost Office Box 1295Birmingham, Alabama 35201Tel 205.992.7872 ENCLOSURE CONTAINSINFORMATION NOT FORPUBLIC DISCLOSURE Fax 205.992.

7601SSOUTHERN NAPR 2 3 2014 COMPANYDocket Nos.: 50-348 NL-14-0446 50-364U. S. Nuclear Regulatory Commission ATTN: Document Control DeskWashington, D. C. 20555-0001 Joseph M. Farley Nuclear PlantResponse to Request for Additional Information Regarding License Amendment Request for Transition to 10 CFR 50.48(c)

-NFPA 805 Performance BasedStandard for Fire Protection for Light Water Reactor Generating PlantsLadies and Gentlemen:

By letter dated September 25, 2012, the Southern Nuclear Operating Company(SNC) submitted a license amendment request (LAR) for Joseph M. Farley Units 1and 2 (Ref. TAC NOS. ME9741 and ME9742).

The proposed amendment requeststhe review and approval for adoption of a new fire protection licensing basis whichcomplies with the requirements in Sections 50.48(a) and 50.48(c) to Title 10 to theCode of Federal Regulations (10 CFR), and the guidance in Regulatory Guide (RG)1.205, Revision 1, Risk-Informed, Performance-Based Fire Protection for ExistingLight-Water Nuclear Power Plants.By letter dated December 12, 2012, the Nuclear Regulatory Commission (NRC)Staff requested supplemental information regarding the acceptance of the licenseamendment (Adams Accession No. ML12345A398).

SNC provided the requested information by letter dated December 20, 2012. The NRC staff subsequently completed the acceptance review by letter dated January 24, 2013, (AdamsAccession No. ML13022A158).

By letter dated July 8, 2013, the NRC Staff formally transmitted a request foradditional information (RAI) related to the referenced license amendment.

SNC'sresponses to these RAIs are being provided by three submittals.

By letter datedSeptember 16, 2013, SNC provided the first set of responses.

By letter datedOctober 30, 2013, SNC provided the second set of responses and by letter datedNovember 12, 2013, SNC provided the remaining set of responses.

SNCprovided that supplemental responses would be provided for nine of the RAIs.

U.S. Nuclear Regulatory Commission NL-14-0446 Page 2By letter dated March 28, 2014, the NRC Staff formally transmitted the secondround of requests for additional information related to the referenced licenseamendment.

The enclosures to this letter provide responses to the RAIs inaccordance with the agreed upon completion dates. The responses to PRA RAI01.01, PRA RAI 06.a.01, and PRA RAI 35 will be provided by May 23, 2014. Asdiscussed, the responses to PRA RAI 01.01, 16.a.01, 21 .a.01, 33.a.01, and33.c.01 have been moved to a May 23, 2014 due date. Attachment G, RecoveryActions Transition, provides a replacement Table G-1. Attachment G containssensitive information and should be withheld from public disclosure under 10 CFR2.390. A revision to Attachment M, License Condition

Changes, is also providedto docket the current proposed version of the license condition.

The No Significant Hazards Consideration determination provided in the originalsubmittal is not altered by the RAI responses provided herein.This letter contains no new NRC commitments.

If you have any questions, pleasecontact Ken McElroy at (205) 992-7369.

Mr. C. R. Pierce states he is Regulatory Affairs Director of Southern NuclearOperating

Company, is authorized to execute this oath on behalf of SouthernNuclear Operating Company and, to the best of his knowledge and belief, thefacts set forth in this letter are true and correct.Respectfully submitted, C. R. PierceRegulatory Affairs DirectorCRP/jkb/lac Sworn o and subscribegi before me this A-3 day of 2014.Notary PublicMy commission expires:

_______/

__

Enclosures:

1. Supplemental Responses to Previous RIA Responses

2. Response to Safe Shutdown Analysis RAI3. Response to Fire Modeling RIAs4. Response to Probabilistic Risk Assessment RAIs U.S. Nuclear Regulatory Commission NL-14-0446 Page 2Attachments:

1. Revision to Recovery Actions Transition-Attachment G2. Revision to License Conditions

-Attachment M3. Revision to Fire PRA Quality-Attachment Vcc: Southern Nuclear Operating CompanyMr. S. E. Kuczynski,

Chairman, President

& CEOMr. D. G. Bost, Executive Vice President

& Chief Nuclear OfficerMs. C. A. Gayheart, Vice President

-FarleyMr. B. L. Ivey, Vice President

-Regulatory AffairsMr. D. R. Madison, Vice President

-Fleet Operations Mr. B. J. Adams, Vice President

-Engineering RTYPE: CFA04.054 U. S. Nuclear Regulatory Commission Mr. V. M. McCree, Regional Administrator Mr. S. A. Williams, NRR Project Manager -FarleyMr. P. K. Niebaum, Senior Resident Inspector

-FarleyMr. J. R. Sowa, Resident Inspector

-FarleyAlabama Department of Public HealthDr. D. E. Williamson, State Health Officer Joseph M. Farley Nuclear PlantResponse to Request for Additional Information Regarding License Amendment Request for Transition to 10 CFR 50.48(c)NFPA 805 Performance Based Standard for Fire Protection for Light WaterReactor Generating PlantsEnclosure 1Supplemental Responses to Previous RAI Responses Provided in the SNC Letter Dated November 12, 2013 Enclosure 1Supplemental Responses to Previous RAI Responses Farley RAI SSA 14Item 6 of Attachment S to the submittal, states that for Fire Area 2-021 aninterposing relay and fuse will be installed to protect cable 2VYDG15J from fireinduced failure and to prevent the breaker from tripping.

It also states that for FireArea 1-021 a fuse will be installed for cable 1VBJ5012F to prevent fire damageand that for Fire Area 2-041 a fuse will be installed for cable 2VAJ5007L toprevent fire damage. Provide clarification that the installation of the relays andfuses is for mitigating the secondary effects of cable damage and not to protectthe cables from fire damage.Response provided by SNC letter NL-13-2269 dated November 12, 2013The modifications identified in this RAI are intended to address common powersupply associated circuits for the Fire PRA, as defined in Section 3 ofNUREG/CR-6850.

The intent of the modifications is to prevent fire damage on theidentified cables from causing a loss of power to circuitry associated withautomatic operation of auxiliary feedwater (AFW). Each specific case isdiscussed below.Additional refinements to the Fire PRA model indicate that it will likely not benecessary to protect Cable 2VYDG15J with an interposing relay and fuse asoriginally described in Attachment S. At present, satisfactory CDF and LERFvalues are achieved without credit for these modifications.

Accordingly, themodifications associated with Cable 2VYDG15J will most likely be removed fromAttachment S of the LAR. Final confirmation of this change will occur when thecomprehensive impact of all model changes is established.

Cable 1VBJ5012F isassociated with automatic operation of Unit 1 auxiliary feedwater (AFW).Depending on the specific functionality credited for affected AFW components, fire-induced failure of this cable can result in a loss of automatic AFW control.

Inmost but not all scenarios the failure does not impact manual control from theMain Control Room. The intent of fusing this cable is to isolate the cable if itsuffers a fire-induced short circuit, thereby preventing a loss of power to theautomatic AFW control circuitry.

The fuse will not prevent hot-short inducedspurious actuation of the automatic AFW control circuit.

For these cases, the FirePRA will reflect the failures, with no credit taken for the modification.

Cable 2VAJ5007L is associated with automatic operation of Unit 2 auxiliary feedwater (AFW). Depending on the specific functionality credited for affectedAFW components, fire-induced failure of this cable can result in a loss ofautomatic AFW control.

In most but not all scenarios the failure does not impactmanual control from the Main Control Room. The intent of fusing this cable is toisolate the cable if it suffers a fire-induced short circuit, thereby preventing a lossof power to the automatic AFW control circuitry.

The fuse will not prevent hot-short induced spurious actuation of the automatic AFW control circuit.

For thesecases, the Fire PRA will reflect the failures, with no credit taken for themodification.

El -1 Enclosure 1Supplemental Responses to Previous RAI Responses Supplemental

Response

After further review of the proposed modification to cable 2VYDG15J and basedon further model refinements completed in response to other RAIs, this plantmodification is no longer required.

The modification to protect cable 2VYDG15Jis no longer credited in the analysis and will be removed from Attachment S.The modifications for cables 1VBJ5012F and 2VAJ5007L are still required toaddress common power supply associated circuits for the Fire PRA. The plantmodification for cables 1VBJ5012F and 2VAJ5007L will be updated inAttachment S based on the findings presented in the original response to RAISSA 14. As identified in the original response to RAI SSA 14, there are onlyspecific function states that the modification will be credited in. An updatedAttachment S will be provided to reflect these changes for the three cablesdiscussed in this RAI response with the response to RAI PRA 35.El -2 Enclosure 1Supplemental Responses to Previous RAI Responses Farley PRA RAI 01(b) through 01(g)In Enclosure 6 to the supplement dated December 20, 2012 (ADAMS Accession No. ML 12359A051),

the results are presented for both the total and delta coredamage frequency (CDF) that are actually lower than previously reported inAttachment W of the LAR. The submittal, although only the credit for the electrical cabinet factor was removed.

With the additional removal of credit for the maincontrol room (MCR) very early warning fire detection system (VEWFDS),

it isexpected that these CDF results would increase, consistent with the increases inthe large early release frequency (LERF) values. Explain why, including any keymodeling assumptions that may be relevant, the increases in total and delta-CDF are now lower than before especially in light of the higher increases in total anddelta-LERF.

From the submittal, address the following:

b. In Section V.2, Sensitivity of Fire PRA Methods, specifically in Tables V.2-2 through V.2-4, discuss whether there are any delta-risk values sensitive to crediting risk reduction from installing VEWFDS in the MCR. If so,discuss how they would change if this credit were removed.c. In Table V-i, Fire PRA Peer Review -Facts and Observations, withrespect to Supporting Requirement (SR) FQ-A3, discuss whether theinstallation of VEWFDS was credited after updating the analysis for theMCR using Appendix L of NUREG/CR-6850, "EPRI/NRC-RES Fire PRAMethodology for Nuclear Power Facilities,"

including how much riskreduction is being realized.

Also, with respect to SR FSS-B2, discuss towhat extent the conclusion of insensitivity is dependent upon this credit.d. In the LAR Table W-1, Summary of Total Plant Risk, and in Enclosure 3 tothe letter dated September 25, 2012, Item 3 on page 2, two of thesensitivity analyses (in LAR Attachment V) increase the CDF by -2E-5/yr, bringing the total close to 1 E-4/yr. Discuss whether all values here, boththese values and the sensitivities, have taken some risk reduction creditfor installing VEWFDS in the MCR. If total CDF becomes >1 E-4/yr or totalLERF > 1 E-5/yr, address any ramifications due to these changes relativeto the guidelines on total risk in RG 1.174, "An Approach for UsingProbabilistic Risk Assessment in Risk-Informed Decisions on Plant-Specific Changes to the Licensing Basis."e. In LAR Attachment W, Table W-2, Fire Initiating Events Individually Representing at least 1% of Calculated CDF for Unit 1, discuss whetheran MCR fire scenario should appear among the dominant ones. Includeconsideration that LAR Table W-6 indicates an MCR abandonment CDFof -3E-7/yr, which could rise to dominate the others, if credit for MCRVEWFDS is not taken.f. For Fire Areas Ul 044 and U2 044, in LAR Table W-6, FNP Fire AreaRisk Summary, discuss why CDF equals LERF and delta-CDF equalsdelta-LERF.

Also, discuss what these values would be without the creditbeing taken for installation of VEWFDS in the MCR.El -3 Enclosure 1Supplemental Responses to Previous RAI Responses

g. In LAR Table W-6, FNP Fire Area Risk Summary, discuss how the totalrisk and delta-risk estimates, including those for RAs, would change ifVEWFDS credit in the MCR were removed.Response provided by SNC letter NL-13-2269 dated November 12, 2013b. The increase in control room risk due to the elimination of credit forincipient detection will be addressed in conjunction with additional refinements of the NUREG/CR-6850, Appendix L credit for the maincontrol board Fire PRA quantification.

c. Credit for VEWFDS was taken prior to and after updating the analysis inthe MCR using the NUREG/CR-6850 Appendix L credit. Elimination ofcredit for incipient detection will be addressed in conjunction withadditional refinements of the NUREG/CR-6850, Appendix L credit for themain control board Fire PRA quantification.

d. The increase in risk for the scenarios in the control room which currently credit VEWFDS will be offset by further refinement of the main controlboard analysis using the guidance of NUREG/CR-6850, Appendix L.e. The updated analysis eliminating the credit for VEWFDS will generateupdated tables including Table W-2 which may now include control roomscenarios.

f. The compliant case for these areas is conservatively set to zero whichresults in a conservative delta risk which is equal to the variant case risk.The analysis used in the LAR submittal was based on application of aCCDP of 0.1 for calculation of CDF. The CCDP value of 0.1 was alsoconservatively applied to the CLERP for calculation of the LERF value.These approaches result in the same LERF to CDF ratio and same delta-LERF to delta-CDF ratio.A revised methodology for calculating the CCDP for the abandonment scenarios is discussed in the response to RAI PRA 33c. Application ofthis approach will differentiate between CDF and LERF. Elimination ofcredit taken for VEWFDS as discussed in the response to item (b) above,will increase only those scenarios crediting VEWFDS. Further refinement of the main control board analysis using the guidance of NUREG/CR-6850, Appendix L will be applied to offset the increase in risk and deltarisk resulting from the elimination of incipient detection credit.g. The change in total risk and delta risk with the elimination of VEWFDScredit will be offset by further refinement of the main control boardanalysis using the guidance of NUREG/CR-6850, Appendix L.El -4 Enclosure 1Supplemental Responses to Previous RAI Responses Supplemental

Response

The revised risk with VEWFDS credit removed will include the composite effect ofthe quantification of other RAI responses and will be provided in the response toPRA RAI 35.El -5 Enclosure 1Supplemental Responses to Previous RAI Responses Farley PRA RAI 01(h)In Enclosure 6 to the supplement dated December 20, 2012 (ADAMS Accession No. ML12359A051),

the results are presented for both the total and delta coredamage frequency (CDF) that are actually lower than previously reported inAttachment W of the LAR. The submittal, although only the credit for theelectrical cabinet factor was removed.

With the additional removal of credit forthe main control room (MCR) very early warning fire detection system(VEWFDS),

it is expected that these CDF results would increase, consistent withthe increases in the large early release frequency (LERF) values. Explain why,including any key modeling assumptions that may be relevant, the increases intotal and delta-CDF are now lower than before especially in light of the higherincreases in total and delta-LERF.

From the submittal, address the following:

h. Calculation SE-C051326701-008, Farley Nuclear Plant, Units 1 and 2,NFPA 805 Fire Risk Evaluations, Version Number 1, dated September 25,2012, was discussed during the site audit. In Tables 2-1a through 2-2b ofthe calculation, Fire PRA Variant Case Results, (Non-) Abandonment Trains A (and B) Alignment, three non-suppression probabilities areassumed -1, 0.1 and 0.02. Discuss the bases for the latter two including their use with respect to crediting VEWFDS installed in-cabinet in theMCR panels.Response provided by SNC letter NL-13-2269 dated November 12, 2013Both non-suppression probabilities (NSP), 0.1 and 0.02, are related to the creditfor VEWFDS. For the instrument rack areas (416 and 471), the 0.02 NSP is usedas presented in Section 13.2 of NUREG/CR-6850 Supplement

1. These tworooms were considered physically separated from where the operators arephysically located by the MCR panels themselves.

The 0.1 NSP is the assumedvalue applied to the scenarios postulated in the control room (401) which iscontinuously occupied by the operators.

The credit for VEWFDS for the control room will be removed and the related firescenarios will be revisited in conjunction with NUREG/CR-6850, Appendix Lcredit as discussed in the response to RAI PRA 01a.Supplemental

Response

The revised risk with VEWFDS credit removed will include the composite effect ofthe quantification of other RAI responses and will be provided in the response toPRA RAI 35.El -6 Enclosure 1Supplemental Responses to Previous RAI Responses Farley PRA RAI 06(a)Section 10 of NUREG/CR-6850, Supplement 1, states that a sensitivity analysisshould be performed when using the fire ignition frequencies in the Supplement instead of the fire ignition frequencies provided in Table 6-1 of NUREG/CR-6850. Summarize the details and provide the results of the sensitivity analysis ofthe impact on using the Supplement 1 frequencies instead of the Table 6-1frequencies on CDF, LERF, delta (A) CDF, and ALERF for all of those bins thatare characterized by an alpha that is less than or equal to one.Response provided by SNC letter NL-13-2269 dated November 12, 2013Section V.2.2 of the Farley NFPA 805 LAR provides the details of the sensitivity analysis related to the bins that have an alpha that is less than or equal to one.Table V.2-3 provides the results of this sensitivity for CDF, LERF, ACDF, andALERF. This is provided for Unit 1 only, however based on the similarities between the two units and similar baseline results it is judged that the results ofthis same sensitivity on Unit 2 will be comparable to that of Unit 1.This sensitivity analysis will be updated to reflect the final baseline risk and will besubmitted via a supplemental RAI response or in response to a potential 2ndround RAI, as appropriate.

Supplemental

Response

The sensitivity analysis will include the composite effect of the quantification ofother RAI responses and will be provided in the response to PRA RAI 35.El -7 Enclosure 1Supplemental Responses to Previous RAI Responses Farley PRA RAI 06(b)Calculation PRA-BC-F-1 1-017, Joseph M. Farley Nuclear Plant, Units 1 and 2,Farley Fire PRA Summary Report, Version Number 1, dated September 14,2012, was discussed during the site audit. With respect to p. D-3, Table D-1,Uncertainty and Sensitivity Matrix: Regarding Task 8, discuss the uncertainty analysis performed based on the updated fire frequencies from Supplement 1 ofNUREG/CR-6850.

Discuss whether this went beyond just performing thesensitivity evaluation in part (a), i.e., discuss whether a parametric uncertainty evaluation was performed using distributional parameters for each binrepresented in the CDF and LERF. Summarize the details and report the results.Response provided by SNC letter NL-13-2269 dated November 12, 2013The uncertainty analysis that was completed and presented in Appendix D ofPRA-BC-F-1 1-017 includes the parametric uncertainty based on the use of theignition frequencies from Supplement 1 of NUREG/CR-6850.

The sensitivity analysis will be updated in conjunction with the update of various Fire PRAanalyses to address RAI responses.

The results will be provided via asupplement to the RAI responses or in conjunction with responses to potential 2ndround RAIs, as appropriate.

Supplemental

Response

The sensitivity analysis will include the composite effect of the quantification ofother RAI responses and will be provided in the response to PRA RAI 35.El -8 Enclosure 1Supplemental Responses to Previous RAI Responses Farley PRA RAI 06(c)On page D-17, Section D.3.1, Fire Ignition Bin Frequencies, of the Calculation PRA-BC-F-1 1-017 (as referenced above), although it is recognized that thesensitivity analyses performed here were done on an earlier FPRA model, onewould still expect the ratio of the resultant CDFs in this table (8.43E-5/4.66E-5

=1.81) to be roughly the same or even slightly lower than the corresponding ratioreported in Table V.2-3 of the LAR, namely 7.46E-5/5.24E-5

= 1.42. Similarly thedelta-CDFs should be roughly the same, with the value in Table V.2-3 (2.22E-5/yr) perhaps slightly larger than the one here (3.77E-5/yr).

Neither case is true.It is recognized that the two sensitivity evaluations may be different, namely theone here addressed ALL the bin frequencies while that in Table V.2-3 addressed only the selected bins with alpha parameters of 1.0 or less. If that is theexplanation, extract the results from here, with appropriate adjustments to reflectthe final FPRA model, and include in the LAR with a discussion of the difference between the two sensitivity analyses.

If there is another explanation, provide it.Response provided by SNC letter NL-13-2269 dated November 12, 2013The ignition frequency sensitivity provided in PRA-BC-F-1 1-017 Appendix D wascompleted on a previous model to the LAR submitted model. The resultsprovided in Appendix D were also based on a sensitivity for all ignition frequency bins, not just those that have an alpha value of less than or equal to 1 inSupplement 1 of NUREG/CR-6850.

These frequencies were also not Bayesianupdated like those used in the base model or the sensitivity in the LAR. It is forthis reason that the two sensitivities are not comparable.

It should be noted thatthe sensitivity provided in LAR Attachment V included the Bayesian updatedignition frequency bins that had an alpha of less than or equal to 1. An updatedsensitivity analysis following the guidance of NUREG/CR-6850 Supplement 1 willbe provided in conjunction with the final baseline model results incorporating theresolution of all RAIs.Supplemental

Response

The sensitivity analysis will include the composite effect of the quantification ofother RAI responses and will be provided in the response to PRA RAI 35.El -9 Enclosure 1Supplemental Responses to Previous RAI Responses Farley PRA RAI 08(a)It was stated at the 2010 industry fire forum that the Phenomena Identification and Ranking Table (PIRT) Panel being conducted for the circuit failure tests fromthe DESIREE-FIRE and CAROL-FIRE tests may be eliminating the credit forControl Power Transformers (CPTs) (about a factor 2 reduction) currently allowedby Tables 10-1 and 10-3 of NUREG/CR-6850, Vol. 2. Provide the results of asensitivity analysis that removes this CPT credit from the PRA, showing theimpact of this potential change on CDF, LERF, ACDF, and ALERF.Response provided by SNC letter NL-13-2269 dated November 12, 2013A sensitivity analysis was performed as part of the Fire PRA development thatincluded doubling all of the hot short probabilities, not just those associated withCPTs. This sensitivity is included in Section V.2.3 of the Farley NFPA 805 LAR,Table V.2-4. The results are provided here. This sensitivity was performed on aversion of the model that preceded the model presented in the LAR and onlyusing the Unit 1 Train A model. This is representative of the other unit and trainsin the Farley analysis based on the type of changes that took place between thetwo models.Table V.2-4Control Power Transformer Sensitivity Case Resultant Delta CDF ChangeCDFCDF -Base 4.66E-05

......CDF -Double HS 4.92E-05 2.60E-06 5.28%ProbInterim guidance has since been accepted by the NRC for the treatment of hotshorts in the Fire PRA. This interim guidance is documented in Interim Technical Guidance on Fire-Induced Circuit Failure Mode Likelihood

Analysis, dated June14, 2013 (ML13165A209, ML13165A214).

These new probabilities also suggestthat the sensitivities presented here will be bounding with respect to theanticipated change in hot short probability guidance.

This is based on theprobabilities found in the interim guidance (which do not provide distinction on thepresence of a CPT) actually being smaller than those found in NUREG/CR-6850 Table 10-2 and 10-4. The sensitivity above is based on all hot short probabilities from NUREG/CR-6850 Table 10-1 through 10-4 being doubled, while the newinterim guidance, even for the CPT circuits, is less than what the probabilities would be if doubled.

These new probabilities also suggest that the sensitivities presented here support the conclusions that the model is not sensitive to thesechanges.

An updated sensitivity analysis using bounding values for hot shortprobability credit will be included in the final baseline Fire PRA quantification.

This will beprovided via a supplemental RAI submittal or in conjunction with apotential 2° round of RAIs, as appropriate.

El -10 Enclosure 1Supplemental Responses to Previous RAI Responses Supplemental

Response

The updated sensitivity analysis will include the composite effect of thequantification of other RAI responses and will be provided in the response to PRARAI 35.El -11 Enclosure 1Supplemental Responses to Previous RAI Responses Farley PRA RAI 08(b)Calculation PRA-BC-F-1 1-017, Joseph M. Farley Nuclear Plant, Units 1 and 2,Farley Fire PRA Summary Report, Version Number 1, dated September 14,2012, was discussed during the site audit. In light of part (a), update theestimates from the table on page D-17, section D.3.2, Spurious Operation Probabilities, using the latest FPRA model, including an update to Table V.2-4 ofthe LAR.Response provided by SNC letter NL-13-2269 dated November 12, 2013Since the submittal of the Farley NFPA 805 LAR, interim guidance has beenaccepted by the NRC for the treatment of hot shorts in the Fire PRA. This interimguidance is documented in Interim Technical Guidance on Fire-Induced CircuitFailure Mode Likelihood

Analysis, dated June 14, 2013 (ML13165A209, ML13165A214).

An updated sensitivity analysis using bounding values for hotshort probability credit (based on latest guidance available at the time) will beincluded in the final baseline Fire PRA quantification.

This will be provided via asupplemental RAI submittal or in conjunction with a potential 2nd round of RAIs,as appropriate.

Supplemental

Response

The sensitivity analysis will include the composite effect of the quantification ofother RAI responses and will be provided in the response to PRA RAI 35.El -12 Enclosure 1Supplemental Responses to Previous RAI Responses Farley PRA RAI 17(b)For Calculation PRA-BC-F-1 1-014, Joseph M. Farley Nuclear Plant, Units 1 & 2,Fire Scenario Development, Version Number 2, dated September 14, 2012,address the following:

a. On pages 14-1 and 14-2, Section 14.0, "Use of Generic Fire ModelingTreatments vs. Detailed Fire Modeling,"

potential conservatisms present in theHughes Approach are credited as a basis for not performing detailed firemodeling.

There appears to be reliance on information available from the EPRIFire Events Database, complete only through 2000 and currently beingupdated, as justification.

Discuss the potential impact on this justification inlight of the fact that such data may be incomplete and, therefore, non-conservatively adapted outside the consensus approach of NUREG/CR-6850.

Response provided by SNC letter NL-13-2269 dated November 12, 2013The discussion cited above is associated with the use of the panel split fractions for defining the fraction of the ignition frequency impacting only the ignitionsource panel with no impact to external targets.

Use of this method was notaccepted by the NRC. A sensitivity evaluation has been performed to addressthe elimination of this factor. The quantification of this sensitivity evaluation willbe incorporated into the baseline fire quantification.

Supplemental

Response

The sensitivity analysis will include the composite effect of the quantification ofother RAI responses and will be provided in the response to PRA RAI 35.El -13 Joseph M. Farley Nuclear PlantResponse to Request for Additional Information Regarding License Amendment Request for Transition to 10 CFR 50.48(c)NFPA 805 Performance Based Standard for Fire Protection for Light WaterReactor Generating PlantsEnclosure 2Response to Safe Shutdown Analysis RAI Enclosure 2Response to Safe Shutdown Analysis RAIFarley Safe Shutdown Analysis (SSA) Request for Additional Information (RAI) 10.01In a letter dated October 30, 2013 (Agencywide Documents Access andManagement System (ADAMS) Accession No. ML13305A105),

the licenseeresponded to SSA RAI 10 and indicated that no recovery actions (RAs) wereomitted from license amendment request (LAR) Attachment G, and that thecorrelation of RAs to variances from deterministic requirements (VFDRs) wereprovided in LAR Attachment G, Table G-1, of the LAR supplement datedDecember 20, 2012 (ADAMS Accession No. ML12359A050).

However, the NRC staff is providing the following examples of inconsistences between:

LAR Attachment C, Table C-1 and LAR Attachment G, Table G-1; withthe LAR supplement and the RAI responses:

a. The LAR Attachment G, Table G-1 submitted on December 20,2012, is missing many fire areas compared to the original LARAttachment G, Table G-1 provided in September, 2012 (e.g., mostof U2-040, U2 2-040, U2 2-041, U2 2-075, and U2 2-076...

up toU2-2-021).

b. Components identified in the new LAR Attachment G, Table G-1don't correspond to VFDRs in LAR Attachment C, Table C-1 (e.g.,Q1P16V0530 and Q1P16V0593).

c. Components corresponding to VFDRs in the LAR Attachment C,Table C-1 are not identified in the new LAR Attachment G, TableG-1 (e.g., VFDRs: U1-044-PCS-040 and U1-1-040-PCS-186),

andthere are potential duplicate VFDRs (e.g., U1-1-040-PCS-145

&146, U1-044-PCS-127

& 128 and U2-044-PCS-079

& 155)d. A partial LAR Attachment G, Table G-1 was submitted with theRAI responses on November 12, 2013 (ADAMS Accession No.ML13318A027),

which includes LAR pages G-9 through G-26. This table has corrections and multiple entries needing to beremoved as duplicates.

However, this LAR Attachment G, TableG-1 still includes several components (e.g., OP-RECOV-XXXX) that are not in LAR Attachment C, Table C-1.Provide LAR Attachment G, Table G-1 and LAR Attachment C, Table C-1 that areup-to-date and correlate accordingly.

RESPONSE

a. It was determined that in the previous transmittal portions of Attachment Gwere inadvertently omitted.

A complete replacement to Attachment G isprovided, in its entirety, as an attachment to this response.

The designinput is provided in the Fire Risk Evaluation document (SE-C051326701-008) where VFDR, component and categorization of DID or risk recoveryaction are documented.

No recovery actions are omitted in Attachment G.No changes to LAR Attachment C are required to incorporate technical E2- 1 Enclosure 2Response to Safe Shutdown Analysis RAIupdates associated with the new Attachment G. Minor administrative updates will be incorporated as an update to the source document(s).

b. The component(s) identified in LAR Attachment G, Table G-1 arecomponent(s) made available by application of recovery action(s).

Incases where a VFDR requires a recovery action (to satisfy the risk ordefense-in-depth criteria of NFPA 805, Ch. 4.2.4) alternate or redundant components are restored to meet the performance criteria.

Theserecovered components are identified in Attachment G, Table G-1. Thecomponents associated with the VFDR are documented in Attachment C,Table C-1. The previous LAR Attachment G tables that were submitted did not include the recovery action type to VFDR to component correlation;

however, to facilitate more complete mapping this has beenrevised with the replacement Attachment G included in this response.

c. See part b (above) to address the scope of response related toAttachment C and Attachment G component differences.

The examples ofduplication have been investigated and corrected where appropriate (U1-1-040-PCS-145 was previously removed).

In some cases there areseemingly duplicate recovery actions against the same VFDR; however,the recovery is associated with a unique basic event with a different action. There are slight differences in the function states or failure modesof the same components that can be easily missed on inspection or mayrequire review of the source documents.

See examples below forillustration of these scenarios:

BE Differences Driving Unique Recovery Actions:SE-C051326701-008 Attachment

-FRE for Unit I Fire Area 1-0092.2.4 Required Modifications The following modifications are required as a result of the risk evaluation for hiis fire area:.Install new trip device in panel Q1R42BOO01B, breaker LB072.2.5 Required Recovery ActionsThe following recovery actions are required as a result of the risk evaluation for this fire area:Table 2-4Required Recovery ActionsComponent ID Basic Event VFDR BE Description Recovery ActionCompnentDescription O1R42BO1B.

EN6R ~OPERATOR RECOVERY Operator action toGIZED:ENERGRIZEDN P-RECOVo U1-1109-OF BATTERY CHARGER provide alternate cooling60EZ 1RCBC-B HVAC-O01 1B OR 2B ROOM for battery charger roomCOOLING FAILS due to loss of fanOPERATOR RECOVERY Operator action to01R42B0001B:ENER OP-RECOV-U1-1-009-OF BATTERY CHARGER provide alternate coolingGIZED:ENERGIZED-IBCSW-B HVAC-001 B ROOM COOLING FAILS- for battery charger dueBC1B SW to the loss of Train BSWE2 -2 Enclosure 2Response to Safe Shutdown Analysis RAIFunction State Differences:

VFDR ID Ul-044-PCS-127 VFDR Q1R43EO001B:AVAILABLE:AVAILABLE-SEQiSHED.

SEQUENCER BUS 1G -QIR43EOOOIB

-SEQUENCER BUS 1G. This normany avaeiable, requiredavailable sequencer.

The electrical system is required to operate for various system supports.

Fire induced damage to cables in the control room and nocontrols on te POS prevent ability to control electrical system components, and a chalenge to lhe Vital Asetaries Nuclear Safety Performance Criteria.

Disposition TiNs condition was evalualed for compliance using the performance-based approach of NFPA 80.S Section 4.2.4. A fire risk evaluation determined that recovevyactions(s) are required to meet applicable risk and DID criteria.

VFDR ID U-0-44-PCS-128 VFDR O1R43E000IBAV.AILABLERAVAILABLE-SHED SEQUENCER BUS IG -O1R43E0001B

-SEQUENCER BUS 1G. This normaty avatilable, requrred available sequencer.

The elec ncal system is r d to operate toe various system supports.

Fire rnduced damage to cables in the control room and no controls on thePCS prevent ability to control electrical system components, and a challenge to the Vital Aunliaries Nuclear Safety Performance Criteria-Disposition This condition was evaluated for compliarce using the p-rormance-based approach of NF PA 805, Section 4.2.4. A fire risk evaluation determined that recoveryactrons(s) are required to meet applicable risk and DID criteria.

Failure Mode Differences:

VFDR ID U2-044-PCS-079 VFDR Q2B41POOO1A:ON:OFF.

RCP 2A -Q2B41POODIA-RCP 2A. This noramally on, required off pump. The RCPs are required off to remove heat generated byrunning pumps added to RCS and Oroti RCS inventory losses thri seals. Fire induced damage to cables and loss o1 dc control power in the contol room and nocntirols on the PCS prevent ability to control pumps, and a chatlenge to the Inventory and Pressure Control Nuclear Safety Performance Criteria.

This condition represents a variance from the deLrministc requirements of Section 4.2.3 of NFPA 805. This is a Separation Issue. Evaluate fon compliance using theperformance-based approach of NFPA 805, Section 4.2.4.Disposition This condition was evablaled for conmplance using the pertormance-based approach of NFPA 805, Section 4.2.4. A tire risk evaluation determined that recoveryactions(s) are reguired to Meet applicable risk and DID cemita.AFFECTED BEIGATE:

U2 B31/B41-RCPPCA-ARC LOSS REASON: Loss of control Loos of powerVFDR ID U2-44-PCS-155 VFDR Q2B41POO1A:ON:OFF.

RCP 2.A- 02B41POOI1A

-RCP 2A. This normally on, required off pump. The RCPs are required of4 to remove hIeat geneated byrunning pumps added to RCS and bimit RCS inventory losses thin seals. Fire induced damage to cables and loss of dc control power in the control room and nocontrols on the PCS prevent abiity to control pumps, and a challenge to the inventory and Pressure Control Nuclear Safety Performance Criteria.

This condition represents a variance from the deterministic reqrirements of Section 4.2.3 of NFPA 805. This is a Separation Issue. Evaiuate for compliance using theperformance-based approach 04 NFPA 805, Section 4.2.4.AFFECTED BEIGATE:

U2 TRIP-RCP-ARC LOSS REASON: Loss of control Loss of powerDisposition This condition was evaluated tor compliance using the performance-based approach o0 NFPA 805. Section 4.2.4. A fire risk evaluation determined that recoveryactions(s) are required to meet applicable 0io and DID criteria.

d. See parts a, b and c for response to this question.

E2 -3 Joseph M. Farley Nuclear PlantResponse to Request for Additional Information Regarding License Amendment Request for Transition to 10 CFR 50.48(c)NFPA 805 Performance Based Standard for Fire Protection for Light WaterReactor Generating PlantsEnclosure 3Response to Fire Modeling RAIs Enclosure 3Response to Fire Modeling RAIsFarley Fire Modeling (FM) RAI 01.02In a letter dated November 12, 2013 (ADAMS Accession No. ML1 3318A027),

thelicensee responded to FM RAI 01 .m and stated:"In certain areas of the plant the physical location is such that there isinsufficient space for placement of a transient fire that would be equivalent to that of the 317 kW transient fire as described in NUREG/CR-6850.

Forexample, a hallway that may be approximately 3 feet wide would notphysically allow a 317 kW fire as defined in the Generic Fire Treatments, Supplement

3. The diameter of 317 kW fire is 1.1 meters (3.6 feet). A fireof 69 kW would be associated with a fire diameter of approximately 1.5feet which is the largest obstruction that is expected to be located in awalkway without obstructing access. Therefore, the lower heat releaserate is used as a more appropriate fire size for the reduced spaceconfiguration."

Provide the following information for the: "certain areas of the plant the physicallocation is such that there is insufficient space for placement of a transient fire..."discussed above:a. A list of the physical locations where a 69 kW transient fire waspostulated based on space limitations.

b. Technical justification for each location why a 317 kW transient firewith physical dimensions that fit within the space could not bepostulated.

c. A discussion of whether the Generic Fire Modeling Treatments (GFMTs) address transient configurations that are small enough tofit within the space (e.g., a hallway) without obstructing access,and that could generate a heat release rate (HRR) higher than 69kW.d. If as a result of the response to items 2 or 3 above, locations areidentified where a transient fire HRR of higher than 69 kW shouldhave been postulated, provide a quantitative assessment of theeffect of the higher HRR on plant risk (core damage frequency (CDF), delta (A) CDF, large early release frequency (LERF), and ALERF).RESPONSE:

a. The Table below provides a list of all rooms where a 69kW transient fire ispostulated.

A short description is also provided for why the smaller(69kW) transient fire is postulated.

Valve Approximately 6'wide, L-shaped room with valves0152 Compartment located along the wall.RoomE3- 1 Enclosure 3Response to Fire Modeling RAIs0202Communication Roomsmall pathway in room along East and North wall,remaining portion of room is taken up by equipment mounted on the ground and wallapproximately 8' wide hallway, multiple doorways into/out0210 Corridor of the hallway with piping located along the wall at thelower elevations 0211 Corridor approximately 4' wide hallway, multiple doorways into/outof the hallway0213 Battery Service approximately 117 sq ft room, multiple doorways into/outRoom of the room, one on each wallapproximately 9' wide, 963 sq ft, a significant portion ofthe room is taken up by trays and conduits that are0227 Cable Chase located in the room, from floor to ceiling, there is minimalfloor space available for transient combustibles to belocatedapproximately 4' wide hallway, multiple doorways into/out0228 Corridor of the hallway, enclosed HVAC located in the middle ofthe room traversing vertically from floor to ceilingapproximately 9' wide, 945 sq ft, a significant portion ofthe room is taken up by trays and conduits that are0300 Cable Chase located in the room, from floor to ceiling, there is minimalfloor space available for transient combustibles to belocated0319 Corridor approximately 4' wide hallway, multiple doorways into/outof the hallwayapproximately 4' wide hallway, multiple doorways into/out0339 Corridor of the hallway, enclosed HVAC located in the middle ofthe room traversing vertically from floor to ceilingapproximately 256 sq ft room, two doorways into/out of0345 Hallway the room, main portion is directly in front of the elevatorand stairwell exit0465 Cable Chase approximately 9' wide, 270 sq ft, this room is onlyaccessible via a ladder from room 04660466 Cable Chase approximately 9' wide, 619 sq ft, this room is onlyaccessible via a ladder from room 0300approximately 9' wide, 945 sq ft, this room is onlyaccessible via a ladder from room 0466, a significant 0500 Cable Chase portion of the room is taken up by trays and conduits thatare located in the room, from floor to ceiling, there isminimal floor space available for transient combustibles to be locatedHot Shutdown approximately 326 sq ft room, significant portion of thePanel Room room is taken up by equipment mounted to the floorE3 -2 Enclosure 3Response to Fire Modeling RAIsapproximately 9' wide, 570 sq ft, a significant portion ofthe room is taken up by trays and conduits that are.hase located in the room, from floor to ceiling, there is minimalfloor space available for transient combustibles to belocated2227Cable C2319 Corridor approximately 4' wide hallway, multiple doorways into/outof the hallwayapproximately 4' wide hallway, multiple doorways into/out2339 Corridor of the hallway, enclosed HVAC located in the middle ofI_ , the room traversing vertically from floor to ceilingb. The transient fire HRR values provided in NUREG/CR-6850 are based ona series of laboratory tests. A review of the tests that are the bases of theNUREG/CR-6850 recommendations found that they all had heatintensities of between 300 and 400 kW/m2.The corresponding

'footprint' of the postulated transient fire that would have an HRR equal to the 98thpercentile value would therefore have an area of approximately 1 M2.Thelocations identified above, walkways and very small rooms, preclude thestorage of material with more than a 1.5 ft2 which is consistent with a 69kW HRR transient fire. Therefore it is not considered credible for a largertransient fire to occur in these locations.

c. Supplement 3, Section 5 to the Generic Fire Modeling Treatments addresses the reduced transient fire size of 69 kW fire. The datapresented in this supplement is representative of the types of transient fires discussed in this response.

The 3' X 3' fuel package fits in all areaswhere a fire greater than 69 kW is postulated.

d. There were no locations where a larger fire was required as a responseto items 2 or 3. Therefore, no qualitative assessment is needed.E3 -3 Enclosure 3Response to Fire Modeling RAIsFarley FM RAI 01.03In a letter dated November 12, 2013 (ADAMS Accession No. ML13318A027),

thelicensee responded to FM RAI 01.p and stated:"The transformers cited in the RAI are filled with Dow Corning 561Transformer Fluid, which is a dimethyl silicone insulating material forpower transformers that has a substantially reduced fire hazard potential than mineral oil insulating materials.

The fluid has a Heat Release Rate(HRR) of 140 kW/m2 per ASTME 1354-90, which is approximately 11times less than that of mineral oil. Based on the burning and ignitioncharacteristics of the Dow Corning 561 Transformer Fluid observed in thepool fire tests (i.e., it is difficult to ignite, it produces short flame heights,and it self-extinguishes),

the transformers containing silicone areconsidered to be similar to a dry transformer ignition source rather than amineral oil filled transformer."

Sections of NUREG/CR-6850, "EPRI/NRC-RES Fire PRA Methodology forNuclear Power Facilities" (such as Chapter 8, Table 8-1 ID 23b, Chapter 11,Table 11-1), indicate that the severity factor of a motor fire can be used tocharacterize a dry transformer.

However, depending

'on the area of the spill, theHRR due to a pool fire from these transformers could be larger than a standalone motor fire.a. Describe and provide technical justification for the HRR andassumed fire source area that were used to characterize transformers filled with the Dow Corning 561 Transformer Fluid.b. Explain how this may impact the plant risk, if the developed firegrows larger than what was originally assumed (i.e., area greaterthan 0.5 m2 for the stated HRR per unit area).RESPONSE:

Part aThe technical justification for the assumed heat release rate is based on theproperties of the Dow Corning 561 silicone liquid as described in the response toRAI FMOD 01 (p). There are three key material properties that substantially reduce the hazard of the silicone liquid relative to conventional hydrocarbon transformer oils. The key properties that are applicable to the transformer ignitionsources are as follows (see "Fact Finding Report on Flammability of LessFlammable Liquid Transformer Fluids -Project No. 87NK1 7807 and Buch, R. R.,Rates of Heat Release and Related Fire Parameters for Silicones, Fire SafetyJournal, Volume 17, 1991):1. The silicone liquid has a fire point that is greater than 3400C (6440F).

Thisattribute suggests that the silicone liquid is difficult to ignite and requireshigh flame heat fluxes to sustain combustion.

2. The silicone liquid forms a silicon dioxide crust during combustion thatcauses the fire to self-extinguish.

This property limits the duration of aE3 -4 Enclosure 3Response to Fire Modeling RAIspotential fire involving liquid spill and may prevent a fire from consuming the full inventory of the liquid spill.3. The heat release rate per unit area of the silicone liquid is 140 kW/m2(12.3 Btu/s-ft2), which is about ten percent the heat release rate per unitarea for conventional hydrocarbon transformer oils per NUREG-1805.

This property indicates that the flames that may develop on ignited liquidwill have a lower power output, which reduces the potential for theformation of a hot gas layer, reduces the flame and thermal plumeexposure to targets located above the burning liquid, and reduces theenergy available for pyrolizing and sustaining combustion at the liquidsurface relative to a conventional hydrocarbon transformer oil.Although the available test data on the silicone liquids indicate that the material iscapable of supporting a fire when a sufficiently large ignition source is applied,there is a precedent for qualitatively crediting the reduced fire hazard potential forindoor transformer installations at commercial nuclear power plants. A recentexample involves an exemption request for the use of operator manual actions infour fire areas at a commercial nuclear power plant (see ML1 10691282 andML1 10700150).

The silicone oil is the predominant combustible in one of the fireareas and a significant fuel load in two other fire areas. The basis for theexemption involves the relatively low fire hazard of the silicone fuel and inparticular the three key material properties previously cited.The basis for linking the heat release rate, and thus the zone of influence (ZOI),to that of a dry transformer with a 98th percentile peak heat release rate of 69 kW(65 Btu/s) is that the silicone liquid would not contribute significantly to the firehazards based on the key material properties cited above. Essentially, it ispostulated that a fire at the silicone liquid transformers would be generally confined to the transformer itself, which is a comparable event to a drytransformer fire.A review of technical specifications and fire protection guidance for silicone liquidtransformers provides qualitative support for this treatment:

Factory Mutual Data Sheet (FMDS) 5-4 addresses fire protection requirements for transformers with the intent of minimizing propertylosses. Per Section 2.2.1.8 and Tables 4 and 5 of FMDS 5-4, indoortransformers with factory mutual approved transformer liquids in factorymutual approved transformers may be located within 0.9 m (3 ft) ofcombustible walls. Given the range of potential combustible wallmaterials, this suggests that the horizontal ZOI for the transformers isexpected to be on the order of 0.9 m (3 ft), which is comparable to the ZOIdimensions for the 69 kW (65 Btu/s) dry transformer.

A factory mutualapproved transformer liquid is one in which the fire point is greater than300'C (572°F),

which the Dow Corning 561 silicone liquid meets.* International Standard IEC 60695-1-40 "Fire Hazard Testing:

Guidance forAssessing the Fire Hazard of Electro Technical Products

-Insulating Liquids,"

Section 7.1 states that there are no reported pool fire incidents involving transformers containing Class K liquids based on 150,000(European) transformers with Class K liquids that are in service.

PerSection 4, a Class K liquid is one in which the fire point is greater than300'C (572'F),

which the Dow Corning 561 silicone liquid meets.E3 -5 Enclosure 3Response to Fire Modeling RAIsBased on the above data, a fire scenario at any of the transformers at the FarleyNuclear Plant containing the Dow Corning 561 silicone liquid is expected to beconfined to the transformer itself rather than result in a spreading pool fire. Themost appropriate ignition source bin for this type of fire scenario is the dry-typeindoor transformer, which per NUREG/CR-6850 has a 98th percentile peak heatrelease rate of 69 kW (65 Btu/s).Part bIf the fire grows larger than originally postulated, the ZOI would increase and thehot gas layer temperature would reach threshold values at shorter time intervals.

This is not unique to the transformer ignition source, but applies to any firescenario.

In the case of the silicone liquid transformers, it is postulated that thefire remains confined to the transformer itself as discussed in Part a of this RAIresponse.

This is based on the properties of the silicone material and on the firehazard guidance provided in IEC 60695-1-40 for Class K insulating liquids.

Assuch, a pool fire having a peak heat release rate per unit area of 140 kW/m2 (12.3Btu/s-ft2) is not associated with this ignition source, regardless of the spread area.E3 -6 Enclosure 3Response to Fire Modeling RAIsFarley FM RAI 01.06In a letter dated September 16, 2013 (ADAMS Accession No. ML1 4038A01 9),the licensee responded to FM RAI 01 .e and stated:"The sensitivity analysis..

.is not directly used in the FPRA. The sensitivity analysis..

.provides an indication of the parameters selections that couldlead to significant variations in the results with the intent that adjustments to the baseline scenarios be made on a case-by-case basis. The currentFPRA uses only the baseline scenarios and therefore does not directlyincorporate the insights provided in the sensitivity analysis of ..." Inaddition, the licensee stated that, "A baseline fire scenario is considered to be conservatively biased if the total probability of control roomabandonment is maximized for the baseline fire scenario.

A baseline firescenario is considered insensitive if the change in the total probability ofcontrol room abandonment remains less than fifteen percent.

A baselinefire scenario is considered to benon-conservatively biased if the change in the total probability of controlroom abandonment exceeds fifteen percent."

Provide technical justification for this 15% limit criterion and explain how it wasdetermined (as opposed to a lower value, such as 5% or 10%).RESPONSE:

The selection of a fifteen percent criterion for determining whether or notuncertainty in a parameter value has a significant effect is based on thetheoretical and observed uncertainty in calorimeter heat release ratemeasurements as described in the SFPE Handbook of Fire Protection Engineering, Section 3-2. The theoretical uncertainty in the heat release ratemeasurements is reported as +/-7 -12 percent, depending on the type of carbonmonoxide correction employed.

The observed uncertainty among test facilities isbetween 17 -23 percent.

Because the heat release rate is a primary inputparameter provided by NUREG/CR-6850, the effect of the uncertainty in theoutput parameters is resolved to a level comparable to that of the heat releaserate input parameter.

When fifteen percent limit criterion is implemented in thecontrol room abandonment parameter sensitivity

analysis, it is applied only to theconservative (upper) limit. A parameter variation is thus allowed to produce anon-conservative result by a factor that exceeds fifteen percent, which implies thebaseline assumption is conservative by more than fifteen percent.

Consequently, the cumulative effect of all parameter variations produces a baselineconfiguration that is conservatively skewed so that the parameter uncertainty isbound.When viewed in context of the overall analysis uncertainty, the parameter uncertainty range of fifteen percent applicable to any given input parameter considered in Attachment 2 of Report 0005-0003-003-001

("Evaluation of ControlRoom Abandonment Times at the Farley Nuclear Power Plant") is significantly narrower than the uncertainty in the probability of control room abandonment attributed to uncertainty in the suppression rate parameter (i.e., A), which is themeans by which the fire PRA assesses the effect of control room abandonment E3 -7 Enclosure 3Response to Fire Modeling RAIstime uncertainty.

This may be shown by comparing the range of control roomabandonment times that could result from parameter uncertainty to the range ofcontrol room times that could result from uncertainty in the suppression rateparameter.

The total probability of control room abandonment for a particular firescenario is determined using the following equation per NUREG/CR-6850:

Pab = Z SFi "Pns, (FM 01.06-1)where Pab is the probability of control room abandonment, SF is the severityfactor for the ignition source heat release rate bin, i is the ignition source heatrelease rate bin number, and Ps,i is the probability that the fire associated withthe ith ignition source heat release rate bin will not be suppressed beforeabandonment occurs (i.e., probability of non-suppression).

For a specific ignitionsource, the severity factor array is constant; thus, the variation in the probability of control room abandonment is entirely reflected in the variation of the probability of non-suppression.

This means that the probability of control room abandonment for each ignition source as a function of time is proportional to probability of non-suppression as a function of time. The probability of non-suppression iscomputed using the following equation per NUREG/CR-6850:

Pns,i = MAX(0.001,exp(-Att))

(FM 01.06-2)where A is a suppression rate parameter and tL is the abandonment time for thefh ignition source heat release rate bin (min). Per NUREG/CR-6850 andNUREG/CR-6850, Supplement 1, the mean value for this parameter is 0.33 min1for fires in the control room however, the fifth percentile value is 0.15 min' andthe ninety-fifth percentile value is 0.58 min-1.Similarly, the suppression rateparameter is equal to 0.126 min-' for transient fire scenarios located outside thecontrol room per NUREG/CR-6850, Supplement 1, with a fifth percentile valueestimated to be 0.086 min" and a ninety-fifth percentile value estimated to be0.166 min1 based on the range provided in Appendix P, Table P-2 ofNUREG/CR-6850.

Finally, the suppression rate parameter is equal to 0.102 min1for electrical panel fire scenarios located outside the control room perNUREG/CR-6850, Supplement 1, with a fifth percentile value estimated to be0.082 min" and a ninety-fifth percentile value estimated to be 0.122 min-' basedon the range provided in Appendix P, Table P-2 of NUREG/CR-6850.

Figure FM01.06-1 graphically depicts the probability of non-suppression for fires in thecontrol room over the 5th -95th percentile uncertainty range for abandonment times between one and twenty-five minutes.

Figure FM 01.06-1 also depicts thefifteen percent variation in the mean value over the same interval.

Figures FM01.06-2 and FM 01.06-3 graphically depict the probability of non-suppression fortransient and electrical panel fires in the electrical equipment area of the controlroom envelope.

In all cases, the fifteen percent variation and the uncertainty inthe suppression rate parameter are constants, therefore the figures of the non-suppression probability are indicative and proportional to of the total probability ofcontrol room abandonment for a given ignition source. Figure FM 01.06-1 showsthat the uncertainty window the MCR parameter sensitivity analysis is resolved tois much narrower than the uncertainty in the probability of control roomabandonment arising from uncertainty in the suppression rate parameter for firesin the control room, which is the means by which the fire PRA assesses the effectof uncertainty in the control room abandonment times. Figures FM 01.06-2 andE3 -8 Enclosure 3Response to Fire Modeling RAIsFM 01.06-3 indicate that the parameter uncertainty is largely bound by theuncertainty in the suppression rate parameter, except for very short times whichdo not arise in the analysis for fires outside the control room. The uncertainty range is less than that for fires in the control room; however, the maximum non-conservative bias for fires outside the control room any parameter considered inAttachment 2 of Report 0005-0030-003-001 is 4.55 percent, which is muchsmaller than the fifteen percent window.Therefore, the basis for the fifteen percent threshold for defining a significant parameter uncertainty effect is the uncertainty in the measured heat release rateinput parameters.

Because the input heat release rates may have a variation of 7-23 percent, a value of fifteen percent is selected as intermediate and typical forlarge scale heat release rate measurements.

The baseline cases are skewed witha conservative factor within this uncertainty range. When viewed in context of theoverall uncertainty in the control room abandonment times, the parameter uncertainty range of fifteen percent applicable to any given input parameter issignificantly narrower than the uncertainty in the probability of control roomabandonment attributed to uncertainty in the suppression rate parameter (i.e., At),which is the means by which the fire PRA assesses the effect of control roomabandonment time uncertainty.

Mean suppression rate parameter (0.33 min-1)5e percentile suppression rate parameter (0.15 min")95th percentile suppression rate parameter (0.58 min')-- -Fifteen percent uncertainty range for parameter sensitivity C0)0.0. 0.1 F0z0.012 0.0010 5 10 15 20 25Control Room Abandonment Time (min)Figure FM 01.06-1 -Probability of Non-Suppression for Control RoomFires (Proportional to the Probability of Control Room Abandonment for aGiven Ignition Source).E3 -9 Enclosure 3Response to Fire Modeling RAIsMean suppression rate parameter (0.126 min"1)5!h percentile suppression rate parameter (estimated, 0.086 min")95th percentile suppression rate parameter (estimated, 0.166 min1)---Fifteen percent uncertainty range for parameter sensitivity 0.a.CLCO0z045IL10.10.010.00105 10 15 20 25Control Room Abandonment Time (min)Figure FM 01.06-2 -Probability of Non-Suppression for Transient Ignition SourceFire Scenarios Located in the Equipment Room (Proportional to the Probability ofControl Room Abandonment for a Given Ignition Source).E3 -10 Enclosure 3Response to Fire Modeling RAIsMean suppression rate parameter (0.102 min")S5h percentile suppression rate parameter (estimated, 0.082 min-)95h percentile suppression rate parameter (estimated, 0.122 min-)--- Fifteen percent uncertainty range for parameter sensitivity 10U)I-C. 0.10zo6 0.010..00.0005 10 15 20 25Control Room Abandonment Time (min)Figure FM 01.06-3 -Probability of Non-Suppression for Electrical Panel FireScenarios Located in the Equipment Room (Proportional) to the Probability ofControl Room Abandonment for a Given Ignition Source).E3 -11 Enclosure 3Response to Fire Modeling RAIsFarley FM RAI 01.07FM RAI 01.1 requested the licensee to assure that non-cable intervening combustibles were not missed and to provide information on how intervening combustibles were identified and accounted for in the fire modeling analyses andthe FREs. In a letter dated November 12, 2013 (ADAMS Accession No.ML13318A027),

the licensee responded to FM RAI 01.1 and explained that duringadditional walkdowns, non-cable intervening combustibles have been identified, and that the quantities of the combustibles (primarily insulation materials) islimited and does not impact the current scenario quantification.

The licenseefurther stated that it is anticipated that the balance of walkdowns will yield similarresults and that the results from the balance of the intervening combustible walkdowns will be assessed upon completion and the impact will be incorporated into the analysis in conjunction with the impact of the secondary cablecombustibles addressed under probabilistic risk assessment (PRA) RAI 17.b.The staff has reviewed the licensees responses to PRA RAI 17b and also FM RAI1 .h (which was referenced in the licensees response to PRA RAI 17b) andconcluded that additional information is required to complete the review. Providea list of the fire scenarios with non-cable intervening combustibles and explainhow the contribution of non-cable intervening combustibles was accounted for inthe zone of influence (ZOI) and hot gas layer (HGL) calculations.

Also discussthe impact on plant risk (CFD, A CDF, LERF and A LERF) of the fire scenarios that involve the non-cable intervening combustibles that were identified in thosewalkdowns.

RESPONSE

The walkdowns performed to identify non-cable intervening combustibles, including insulation materials, were done at the same time as the walkdowns toidentify intervening cable combustibles.

These walkdowns did not identify anyfixed non-cable intervening combustibles of a significant quantity/size that werewithin the zone of influence of an ignition source (the zone of influence for cabledamage was used as a conservative zone of influence for ignition of non-cable intervening combustibles).

These walkdowns are documented in report 0005-0030-003-006, Walkdowns:

Expanded Zone of Influence Target Walkdowns, Rev0. Non-cable intervening combustibles that are enclosed within a pump (pumplube oil) or are contained within a metal enclosure (paper inside a filing cabinet)were not included since they would be exposed to limited radiant energy and arenot likely to ignite and therefore would not contribute to an increase in the heatrelease rate for the initiating fire scenario.

Should these materials ignite, they areexpected to have little impact on components/cables outside of their enclosure given the limited supply of air available for the fuel. This configuration would limitthe energy released inside the metal enclosures.

In addition, there is no directmeans of energy transport from the enclosure to the surroundings; it requires aconduction process through the enclosure boundary and convection and thermalradiation process at the enclosure surface.

This restricts further the rate of energytransfer from the ignited contents to the surroundings.

The non-cable intervening combustibles did not impact any existing scenarios orwarrant creation of new scenarios, therefore, the non-cable intervening combustibles did not impact the Fire PRA risk.E3 -12 Enclosure 3Response to Fire Modeling RAIsFarley FM RAI 02.01In letter dated November 12, 2013 (ADAMS Accession No. ML13318A027),

thelicensee responded to FM RAI 02.f and stated:"This approach is supported by the in-process Fire PRA Frequently AskedQuestions (FAQ 13-0004, "Clarifications Regarding Treatment ofSensitive Electronics.")

Walkdowns to identify sensitive electronics components which are located outside of a panel enclosure are inprogress.

Sensitive electronics credited for post fire shutdown which arelocated outside of a panel enclosure will be evaluated with respect topotential damage by ignition sources in its vicinity using the NUREG/CR-6850 Appendix H criteria for solid-state control components."

Fire PRA Frequently Asked Question (FAQ) 13-0004, which has now beenissued, provides a few limitations to applying this methodology:

a. FAQ-13-0004 states cable damage thresholds can be used fortemperature sensitive equipment inside cabinets provided that (i) thesensitive electronic component is not mounted on the surface of thecabinet (front or back wall/door) where it would be directly exposed to theconvective and/or radiant energy of an exposure fire, and (ii) the presenceof louvers or other typical ventilation means do not invalidate the guidanceprovided.

Describe the limitations that were considered in thedetermination of damage condition for sensitive electronic equipment enclosed in cabinets and explain whether they are in accordance with theFAQ or some other method.b. The conclusions of the Fire Dynamic Simulator (FDS) analysis in FAQ-1 3-0004 are based on radiant heat flux exposure to the cabinet.

Therefore, the 650C temperature damage criterion must still be assessed for othertypes of fire exposures to the enclosed sensitive electronics.

Describewhat temperature damage criterion was assessed and whether it is inaccordance with the FAQ or some other method.RESPONSE:

a. Walkdowns were performed for identification of sensitive electronics mounted outside of electrical cabinets.

Sensitive electronic components inside a cabinet mounted on the front or back wall/door of the cabinet arenot considered to be directly exposed to the convective and or radiantenergy of an exposure fire as long as they are provided with a cover orface plate at or near the panel surface which protects the sensitive electronic components within the panel from exposure to convective andor radiant energy. For panels that have louvers that may contain sensitive electronics, the fire induced damage to these would be included in the HotGas Layer scenario.

If the panel were to be located within the Zone ofInfluence of the exposing ignition source, the cabinet would be failedregardless if louvers were present or not. The guidelines of Fire PRAE3 -13 Enclosure 3Response to Fire Modeling RAIsFAQ 13-0004 were used to define the scope and criteria applied duringthe walkdowns.

b. The potential impact of a hot gas layer on remotely located panels (panelsoutside the zone of influence of the ignition source) is addressed in theFarley Fire PRA. The hot gas layer analysis is based on an 80'C hot gaslayer temperature which has been defined as the bounding temperature for validity of the Generic Fire Treatment zone of influence.

A non-suppression probability is applied to the ignition frequency for the hot gaslayer scenario (scenario impacting all targets within an enclosed volume,single fire zone or multiple fire zones with openings between the firezones) based on the time required to reach a hot gas layer temperature of80°C and the non-suppression probability associated with that timeframe (NSP for 80'C hot gas layer). When an 80'C hot gas layer temperature has been exceeded in the compartment all sensitive electronics including all cables and components in the enclosed volume are assumed to bedamaged.

At and below the 80'C hot gas layer temperature, sensitive electronics located inside of a electrical cabinets are considered to beundamaged, given that the difference between the 80'C hot gas layertemperature and the 65°C sensitive electronics damage temperature isaccounted for by the temperature difference between the hot gas layer(which is typically at the upper elevations of the compartment) and thetemperature within the enclosed panel. This temperature difference isexpected to be significantly higher than the 15'C temperature difference between the 80'C criteria and the 65°C sensitive electronics damagecriteria.

This is due to the difference in elevation of the hot gas layerversus the panel internals, the thermal gradient associated with hot gasesin the compartment, convection boundary condition at the surface of thepanel, convection boundary condition inside the panel. All of these factorsresult in a temperature reduction between the cabinet internals and thehot gas layer. This temperature difference is expected to be significantly larger than 15'C which ensures that sensitive electronics located withinthe electrical cabinet are not damaged by the hot gas layer.E3 -14 Enclosure 3Response to Fire Modeling RAIsFarley FM RAI 02.02In letter dated November 12, 2013 (ADAMS Accession No. ML13318A027),

thelicensee responded to FM RAI 02.g and stated:"Sensitive electronics are typically not located outside of panel enclosures as well as enclosures which protect the electronic modules from theimpact of external environments such as dust. As noted in the responseto RAI FM 02(f), walkdowns to identify sensitive electronics are beingperformed to identify electronic components mounted outside of panelenclosures.

Such components relied upon for post fire shutdown will befurther evaluated, as discussed in RAI FM 02.f, to ensure their availability post fire."Provide the results of the additional walkdowns and confirm that the findings, ifany, have been incorporated into the FM and PRA analysis.

RESPONSE

The walkdowns performed to identify exposed sensitive electronics were done atthe same time as the walkdowns to identify intervening cable combustibles.

These walkdowns are documented in report 0006-0030-003-006, Walkdowns:

Expanded Zone of Influence Target Walkdowns, Rev. 0. No sensitive electronics associated with equipment credited in the Fire PRA analysis were located outsideof panel enclosures.

Therefore, the results of these walkdowns had no impact onthe Fire PRA analysis.

E3 -15 Enclosure 3Response to Fire Modeling RAIsFarley FM RAI 07National Fire Protection Association Standard 805 (NFPA 805), "Performance-Based Standard for Fire Protection for Light Water Reactor Electric Generating Plants,"

2001 Edition, Section 2.4.3.3, on acceptability states: 'The PSAapproach,

methods, and data shall be acceptable to the AHJ." The staff hasnoted the utilization of a number of accepted tools and methods in the analysesfor transition such as the Consolidated Model of Fire Growth and SmokeTransport (CFAST) and GFMTs approach.

a. Identify any fire modeling tools and methods that have been usedin the development of the NFPA 805 LAR that are not alreadydocumented in the LAR and where their use or application isdocumented.

Examples might include a methodology (empirical correlations and algebraic models) used to convert damage timesfor targets in Appendix H of NUREG/CR-6850 to percent damageas a function of heat flux and time or supplements to the GFMTs -Empirical Correlations and Algebraic Models.b. For any tool or method identified in "a." above, provide theVerification and Validation (V&V) basis if not already explicitly provided in the LAR (for example in LAR Attachment J).RESPONSE:

a. The elimination of the "panel factors" method for which a sensitivity evaluation was documented in LAR Attachment V,Section V.2.1(Electrical Cabinet Fire Severity Methodology) resulted in additional refinement of the Fire PRA model. The refinements applied included amore detailed evaluation of severity factors with respect to fire impact tothe first external target located outside of an electrical panel. Thisapproach includes the time to damage of the target, which utilizes thedamage delay criteria provided in NUREG/CR-6850 Section H.1.5.2.

Thistime to damage data is incorporated in the evaluation of damage time dueto the impact of a hot gas layer. This approach, which is addressed indetail in the RAI FM 10 response, is an application of the data provided inNUREG/CR-6850 and the Generic Fire Modeling Treatments.

No otherrefinements of significance were applied beyond those alreadydocumented in the LAR.b. No methods requiring Verification and Validation beyond those addressed in the Farley NFPA 805 LAR Attachment J have been used. The use ofthe NUREG/CR-6850, Appendix H data described in the response to item"a" above was verified in the same manner as other Fire PRA supporting calculations.

No validation of this method is required since it is based ondata specified in NUREG/CR-6850.

Further discussions of the basis forthis method are provided in the response to RAI FM 10.E3 -16 Enclosure 3Response to Fire Modeling RAIsFarley FM RAI 08Explain how high energy arcing fault (HEAF) initiated fires were addressed in theHGL and Multi Compartment Analysis (MCA) and provide technical justification for the approach that was used to calculate HGL development timing. Morespecifically, confirm if the guidance provided in NUREG/CR-6850, pages 11-19,fourth bullet regarding the fire growth, and the guidance provided on page M-13,sixth bullet regarding delay to cable tray ignition was followed.

Also, considering the energetic nature of the HEAF event, provide justification for the HRR used inthe HGL calculations for electrical cabinet fires following a HEAF event.RESPONSE:

The HEAF initiated fires are currently addressed in the Farley Fire PRA, including scenarios for HGL and MCA. The current approach uses the Zone of Influence (ZOI) based on NUREG/CR-6850 Appendix M. The heat release rate (HRR) thatis currently used is based on a medium voltage switchgear (MVSG) or LoadCenter (LC) as applicable in conjunction with the HRR associated with anysecondary combustibles that may be part of the applicable ZOI. These HRRs aredefined in Report 0005-0030-003-002, Combined Ignition Source -Cable TrayFire Scenario ZOls for Farley Nuclear Power Plant Applications Rev 1.The current HGL and MCA uses the fire growth of 12 minutes, which isassociated with Electrical Cabinet fires as defined in NUREG/CR-6850 AppendixG.To account for the guidance provided in NUREG/CR-6850 Appendix M, fourthand sixth bullet, the current analysis will be updated to apply the peak HRR to theHEAF scenarios at t=0. The HRR in these cases will include that of the ignitionsource (MVSG or LC) and any applicable secondary combustibles.

This updateto the methodology will be applied to both the HGL and MCA, and the results ofthis incorporation will be provided with the response to RAI PRA 35.E3 -17 Enclosure 3Response to Fire Modeling RAIsFarley FM RAI 09In a letter dated November 12, 2013 (ADAMS Accession No. ML13318A027),

thelicensee responded to FM RAI 01 .h and stated:"In lieu of demonstrating which scenarios are conservative and whichrequire further analysis, new ZOI tables have been developed that areapplicable to ignition source-cable tray configurations at Farley."

Inaddition, the licensee stated that 'The method used to develop the ZOIdimensions includes the vertical cable tray stack propagation modeldescribed in NUREG/CR-6850, Appendix R, the FLASH-CAT calculation method described in NUREG/CR-7010, Volume 1, and the radiant heatflux calculation methodology described in the GFMT document."

Describe the methodology used in the revised analysis to determine the ignitiontime of the first cable tray above an ignition source. If it was assumed that thelowest cable tray in a stack located above an ignition source will not ignite unlessthe tray is located below the flame tip of the ignition source fire, provide technical justification for this assumption and provide the results of a sensitivity analysis todemonstrate that the conservatism of the ZOI and HGL calculations for fires thatinvolve cable trays as secondary combustibles is not adversely affected by theignition criterion that was used (compared to the ignition criteria in NUREG/CR-6850 and NUREG/CR-7010, "Cable Heat Release,

Ignition, and Spread in TrayInstallations During Fire").RESPONSE:

The bottom cable tray in a cable tray stack is ignited one minute after the ignitionsource ignites in the FLASH-CAT calculations used to support the Zone ofInfluence (ZOI) and Hot Gas Layer (HGL) tabulations provided in Report 0005-0030-003-002, Rev. 1 and Report 0005-0030-003-003, Rev. 0 (see Assumption 18 in Report 0005-0030-003-002, Rev. 1, for example).

This corresponds to theminimum damage time for thermoset cable targets listed in Tables H-6 and H-8 ofNUREG/CR-6850 and is more conservative than the generic value of five minutesthat is assumed in NUREG/CR-7010, Volume 1.The physical basis for the assumption that the lowest cable tray in a stack locatedabove an ignition source ignites if it is at or below the flame tip is the full scaletest data available for cable tray ignition and propagation data. There are threebasic test series upon which the empirical flame spread model for cable tray firesas provided in NUREG/CR-6850 and validated in NUREG/CR-7010, Volume 1 isbased. These are as follows:" EPRI-NP-1881 (Sumitra tests);* NUREG/CR-0381 (Klamerus tests); and* NUREG/CR-7010, Volume 1 (NIST tests).The test reports for these test series document the results of about thirty-five toforty open configuration, unprotected cable tray fire tests. In all cases, the initialE3 -18 Enclosure 3Response to Fire Modeling RAIsignition source for the lowest cable tray within a stack is a gas burner or liquid fuelpan fire that causes flame impingement on the lowest cable tray in the stack.Further, there are no cases presented in which the thermal plume above theflame tip alone was sufficient for igniting a cable tray. An indication of this effectmay be observed in the test series presented in NUREG/CR-0381, Test 28,which was a two tray stack with ceramic blanket on the top of both trays. Thepropane burner was sufficiently large to ignite the lower tray and to expose theupper tray to the thermal plume during the exposure fire cycle, but the upper traydid not ignite. A quantitative indication of the conditions necessary for fire ignitionand surface spread is provided in NUREG/CR-5384.

Burn mode evaluations forboth non-rated (thermoplastic) and low flame spread (thermoset) cables arepresented and indicate that for thermoplastic cables, which bound the results forthermoset cables, a surface temperature of 5380C (1,000°F) and an internal fueltemperature of 5770C (1,0700F) are necessary for surface flames to develop (seeFigure FM 09-1). Smoldering and pyrolysis occur at lower temperatures, and adeep seated fire may result if the internal fuel temperature is approximately 5380C (1,000°F) regardless of the surface temperature.

1* 4000SL1200 OJ10 200 0 0 0101010101

-6000400200 *wT00 200 400 600 800100012001400160018 00FUEL NTEF" TEW ('F)Figure 3.5: Burn Mode Analysis of Non-Rated Cable FireFigure FM 09-1 -Burn Mode Analysis of Thermoplastic Cables perNUREG/CR-5384.

E3 -19 Enclosure 3Response to Fire Modeling RAIsIt can be shown using the Heskestad flame height correlation, the Heskestad virtual origin correction, and the Heskestad plume centerline temperature correlation that the temperature at the flame tip is approximately equal to 5280C(9830F), which is lower than the minimum temperature of 5770C (1,0700F)needed for surface flame spread per Figure FM 09-1. The Heskestad flameheight correlation is given as follows per Section 2-1 of the SFPE Handbook ofFire Protection Engineering.

Lf = -1.02D + 0.235Q0°4 (FM 09-1)where Lf is the flame height (m), D is the effective fire diameter (m), and 0 is thetotal heat release rate of the ignition source (kW). The Heskestad plumecenterline temperature correlation is given as follows per Section 2-1 of theSFPE Handbook of Fire Protection Engineering and "Fire Plumes and CeilingJets" in the Fire Safety Journal, Vol. 11, Nos. 1 & 2:T, = T.,o + 22Q02/3(Z -Zo)-513 (FM 09-2)where T, is the plume centerline temperature (0C) at an elevation Z (m) above thefire base, Too is the initial temperature (0C), and zo is the height of the virtual originbelow the fire base (m). The virtual origin height is given by the following equationper Section 2-1 of the SFPE Handbook of Fire Protection Engineering zo = -1.02D + 0.08300.4 (FM 09-3)where all terms have been defined.

At the flame tip, the height above the firebase, Z, is equal to the flame height, Lf. Combining Equations FM 09-1, FM 09-2, and FM 09-3 results in the following:

Tc = T,, +

= Tcx + 508 (FM 09-4)where all terms have been defined.

The plume centerline temperature is thusindependent of both the fire diameter and the heat release rate at the flame tipand is equal to 528°C (983'F) for an ambient temperature of 200C (680F).Ambient air temperatures greater than 690C (1560F) would be necessary tocause the peak plume temperature to exceed the minimum value of 5770C(1,070°F) necessary for surface flames to develop.

There are no plant areas inwhich the ambient air temperature is greater than 690C (156°F);

thus, theassessment is generally applicable.

These calculations are consistent with the observation provided in Section 7.2 ofNUREG/CR-7010, Volume 1 that the damage threshold for cables ascharacterized by a heat flux is not a good indicator of ignition.

Based on the conecalorimeter tests summarized in NUREG/CR-701 0, Volume 1, a heat flux ofexposure of 25 kW/m2 (2.2 Btu/s-ft2) is minimally sufficient to cause ignition andsustained burning for all classes of cables considered, including the thermoplastic cables which bound the results of thermoset cables. Data provided in Section 2-14 of the SFPE Handbook of Fire Protection Engineering indicates that the netheat flux to an object immersed in the fire plume at the flame height asE3 -20 Enclosure 3Response to Fire Modeling RAIsdetermined from the stagnation point is between 5 -15 kW/m2 (0.44 -1.32 Btu/s-ft2), which is significantly less than the minimum heat flux necessary to causesustained ignition per NUREG/CR-7010, Volume 1. This is further supported bythe test data for thermoplastic cables provided in Figures 10 -12 of NUREG/CR-6931, Volume 3. A shroud temperature (exposure temperature) of about 300 -3300C (572 -626-F) is used to heat various types of thermoplastic cables inorder to determine the damage times. Although the focus of the tests was not onthe ignitability of the cables, the temperature profiles provide an indication of thecable response to the temperature exposure.

The figures indicate that the cablesdo not ignite over the ten to twenty minute exposure interval and typically showthe cables reach a steady state temperature close to 3000C (5720F) even thoughdamage via electrical short occurs around 2000C (3920F).The requirement for flames to impinge on the lowest cable tray before ignition isassumed is considered reasonable and supported by the available documents.

Asummary of the basis is as follows:* All cable tray fire test data involves an ignition source that results in flameimpingement on the lower cable tray in a cable tray stack;* A burn mode analysis of thermoplastic cables suggests that the minimumtemperature required for surface flames or deep seated burning todevelop is approximately 5380C (1,000°F) for thermoplastic cables, whichbound the results for thermoset cables; and* Ambient air temperatures greater than 690C (1 560F) would be necessary to cause the peak plume temperature to exceed the minimum value of577°C (1,0700F) necessary for surface flames to develop.

There are noplant areas in which the ambient air temperature is greater than 690C(1560F); thus, the assessment is generally applicable.

Note that the flame height is used as an indicator of ignition and the ZOIdimensions are used as an indicator of damage. Consequently, cable trayslocated above an ignition source may be above the flame tip but still be within theZOI for cable damage.E3 -21 Enclosure 3Response to Fire Modeling RAIsFarley FM RAI 10In letter dated November 12, 2013 (ADAMS Accession No. ML13318A027),

thelicensee responded to PRA RAI 20.c and stated:"The approach discussed in Assumption 7 of the Sensitivity Analysis usesthe data provided in NUREG/CR-6850 Appendix H for time to damage asa function of heat flux to define an accrual of damage based on the time ateach heat flux. The value is taken from the fire model at a given distanceand correlates that to a fraction of the accrued damage by dividing thetime at the heat flux by the time at that heat flux required to causedamage to the cable. Cable damage occurs when the accrued damageequals 1.0. This approach uses the same principles that are applied toequipment qualification of safety related equipment including cables forpost-accident environments, such as inside containment LOCAconditions.

The potential for some non-conservatism arises from the dataspecified in Appendix H where no damage is accrued regardless of thetime exposure when the heat flux is just below the damage heat flux. Toeliminate this potential non-conservatism, the analysis is being updated toassume a bounding damage accrual during the time period prior to thecable reaching the critical heat flux."The methodology assumes that the "damage rate" at a specified heat flux as thereciprocal of the failure time in Tables H-7 (for thermoset cable targets) and H-8(for thermoplastic cable targets) of NUREG/CR-6850.

There does not appear tobe a physical basis for this assumption.

Provide evidence of the validation of themethodology as a whole, and the damage rate assumption in particular.

RESPONSE

The methodology used to evaluate cable percent damage is based on the use ofthe time to damage data provided in Appendix H of NUREG/CR-6850 andapplying an Arrhenius methodology, which is used extensively for environmental qualification (EQ) of components such as cables in a containment accidentenvironment, to determine the time to damage of the cables. An NRC internalevaluation of the Arrhenius methodology for equipment qualification is provided ina February 24, 2000 NRR Memo from Samuel J. Collins to Ashok Thadani(ML003701987).

The NUREG/CR-6850 Appendix H, Table H-7 data provides times to targetdamage for thermoset cables for a set of steady state incident heat flux values.This provides a time delay for target damage beyond the damage heat flux of 11kW/m2.For instance, Table H-7 provides a 19-minute time-to-damage delay for athermoset cable with a steady state incident heat flux of 11 kW/m2.In order to apply the NUREG/CR-6850 Appendix H data to a fire with a t2 growthrate, the EQ methodology of damage accrual is applied.

The times to damageprovided in NUREG/CR-6850 Appendix H were converted to damage rates bytaking the reciprocal of the time to damage. For instance, the 19 minute time todamage for an 11 kW/m2 incident heat flux in Table H-7 is converted to a min-'19damage rate. This provides a discrete set of damage rates for the heat fluxE3 -22 Enclosure 3Response to Fire Modeling RAIsvalues provided in Appendix H. An exponential regression is applied to thesedata points to generate a damage rate -heat flux profile.

This regression analysis provides the Arrhenius curve for these cables based on the NUREG/CR-6850 Appendix H data.The Farley Fire PRA model used a damage rate profile that assumed no damagebefore a critical incident heat flux was reached, directly applying the Appendix Hdata which states that no damage occurs prior to critical heat flux. The updatedmethodology, referred to in the response to RAI PRA 20, referenced in this RAI,updates the model to assume a damage rate equal to the critical heat fluxdamage rate for incident heat flux values up to and including the critical heat flux.This approach bounds any degradation of the cable target for heat flux valuesbelow the critical heat flux. Beyond the critical heat flux, the Arrhenius curvedamage rates are applied with no maximum damage rate applied, making thisapproach more conservative than that defined by the NUREG/CR-6850, Appendix H data. This ensures the use of a bounding damage rate curve withoutextrapolating data to lower heat flux values, using the critical heat flux damagerate as a minimum damage rate, and not imposing damage rate limits beyond amaximum heat flux, thereby providing a conservative, bounding analysis.

Figure 1 below shows a plot of the damage rate -heat flux profile that modelsthis approach.

The figure shows the plot of the damage rate as specified inNUREG/CR-6850, Table H-7 (square and triangular data points) versus thebounding curve as it is being use in the Farley Fire PRA (square and diamonddata points).

The Farley Fire PRA approach ensures conservatism in the rangeof heat flux values where Table H-7 does not provide data points and itconservatively, with respect to NUREG/CR-6850, extrapolates the data for higherheat flux values using the Arrhenius methodology.

The Farley approach usesbounding data points with respect to the data specified in NUREG/CR-6850, Appendix H (specifically for heat flux values below the critical heat flux and fordata points where the damage rate exceeds 1.0/minute).

See Farley Hot Gas Layer and Multi-Compartment Analysis and Scenario Reportfor details of the application of this evaluation to the Fire PRA model.E3 -23 Enclosure 3Response to Fire Modeling RAIsC4-mE01.51* Damage Accrual Approach* NUREG/CR-6850

• both methods5 Expon. (both methods)o0 , -- T ,0 51015 20 25Heat Flux (kW)Figure 1. Damage rate -heat flux profileE3 -24 Joseph M. Farley Nuclear PlantResponse to Request for Additional Information Regarding License Amendment Request for Transition to 10 CFR 50.48(c)NFPA 805 Performance Based Standard for Fire Protection for Light WaterReactor Generating PlantsEnclosure 4Response to Probabilistic Risk Assessment RAIs Enclosure 4Response to Probabilistic Risk Assessment RAIsFarley PRA RAI 01.01LAR Attachment V, Table V.2-2, provides the results of the electrical cabinet fireseverity sensitivity analysis for Unit 1, also indicating similar results for Unit 2.There, the base CDF rose from 5.24E-5/y to 7.05E-5/y, an increase of 1.81 E-5/y.For A CDF, the base value rose from 8.80E-6/y to 1.03E-5/y, an increase of1.50E-6/y.

The analogous results for LERF and A LERF were as follows:

(1) aLERF increase of 2.59E-6/y from 1.26E-6/y to 3.85E-6/y; (2) a A LERF increaseof 9.90E-8/y from 4.14E-7/y to 5.13E-7/y.

Subsequently, the LAR wassupplemented by a sensitivity analysis which included the effect of removingcredit for very early warning fire detection system (VEWFDS) in the main controlroom (MCR) in addition to the electrical cabinet fire severity adjustment.

Theresults were as follows:

(1) CDF now rose only 1.41 E-5/y (vs. the previous 1.81 E-5/y); (2) A CDF now rose only 1.18E-6/y (vs. the previous 1.50E-6/y);

(3) LERFnow rose more by 6.28E-6/y (vs. the previous 2.59E-6/y);

(4) A LERF now rosemore by 2.88E-7/y (vs. the previous 9.90E-8/y).

In a letter dated September 16, 2013 (ADAMS Accession No. ML14038A019),

asjustification for the smaller increase for CDF and A CDF with credit for bothVEWFDS and electrical cabinet severity adjustment

removed, the licenseeindicated via Table 1 that, in addition to removing credit for VEWFDS in the MCR,the following additional refinements were now included:

(1) refined main controlboard (MCB) fire scenarios (via App. L of NUREG/CR-6850);

(2) more realistic probabilities for HGLs; (3) refined circuit analysis for selected fire scenarios; (4)correction to anomalies in fire ignition frequencies for selected fire scenarios.

Asa result, the CDF and A CDF increase for removing both VEWFDS and electrical cabinet factor credit were actually less than prior to removal of the VEWFDScredit alone. While the licensee's explanation is sound for these metrics, itremains unclear as to why the LERF and A LERF increases do not display thesame trend as CDF and ACDF. If the CDF and A CDF showed a smallerincrease with the additional refinements, why did not the LERF and A LERF aswell? Explain why the increases in LERF and A LERF after removal of theVEWFDS credit and addition of the four refinements trended upward vs. thedownward trend for the CDF and A CDF increases.

RESPONSE

The removal of credit for VEWFDS and electrical cabinet factor credit, along withthe four refinements described above is included in the composite effect of thequantification of other RAI responses to be provided in the response to PRA RAI35.E4 -1 Enclosure 4Response to Probabilistic Risk Assessment RAIsFarley PRA RAI 06.a.01In a letter dated November 12, 2013 (ADAMS Accession No. ML13318A027) thelicensee responded to PRA RAI 06(a) and stated that section V.2.2 of the LARprovides the details of the sensitivity analysis related to the bins that have analpha that is less than or equal to one. Indicate if the acceptance guidelines ofRegulatory Guide (RG) 1.174, "An Approach for Using Probabilistic RiskAssessment in Risk- Informed Decisions on Plant-Specific Changes to theLicensing Basis," may be exceeded when this sensitivity study for those binswith an alpha less than or equal to 1 is applied to the integrated study of PRA RAI35 (see below). If these guidelines may be exceeded, provide a description offire protection or other measures that can be taken to provide additional defensein depth (DID) (see FAQ 08-0048).

RESPONSE

The response to this RAI will be provided by May 23, 2014.E4 -2 Enclosure 4Response to Probabilistic Risk Assessment RAIsFarley PRA RAI 16.a.01In a letter dated October 30, 2013 (ADAMS Accession No. ML1 3305A1 05), thelicensee responded to PRA RAI 16.a and partially addressed some of the criteriafor assuming damage within MCR panels to be limited to the initiating panel,namely the presence of no openings and a double wall with an air gap. However,Appendix S of NUREG/CR-6850 also states that there be no sensitive electrical equipment in the adjacent cabinet (or else such equipment to have already been"qualified" above 82C), even with the double wall with air gap. Otherwise damage to such equipment should be postulated.

Explain whether theseadditional criteria are met or not. If the latter, explain how damage is modeled or,if not, the basis for assuming no damage. (Also see PRA RAI 33.a.01.)

RESPONSE

The response to this RAI will be provided by May 23, 2014.E4 -3 Enclosure 4Response to Probabilistic Risk Assessment RAIsFarley PRA RAI 21.a.01In a letter dated September 16, 2013 (ADAMS Accession No. ML14038A019),

the licensee responded to PRA RAI 21 .a and confirmed that the three severityfactors, 5.02E-4, 4.84E-4 and 0.00158, do not derive from Figure L-1 inNUREG/CR-6850 but are specifically calculated based on the type of ignitionsource, scenario location and abandonment time for the MCR abandonment analysis.

The three severity factors correspond to the abandonment probabilities for transient ignition

sources, equipment room fixed ignition sources and MCRfixed ignition

sources, respectively.

Provide a discussion of the derivation ofthese factors, including their bases, e.g., as given in Section 13.2.1 of the FarleyScenario Development Report, PRA-BC-1 1-014, and Section 6 of Units 1 and 2Control Room Abandonment Times at the Joseph M. Farley Nuclear Plant, Rev.0.RESPONSE:

The response to this RAI will be provided by May 23, 2014.E4 -4 Enclosure 4Response to Probabilistic Risk Assessment RAIsFarley PRA RAI 29.01In a letter dated September 16, 2013 (ADAMS Accession No. ML14038A019),

the licensee responded to PRA RAI 29 and indicated that 22 supporting requirements (SRs) fail to meet Capability Category (CC) II, 17 more than thestaff was able to determine by review of LAR Attachment V, Table V-1. Thelicensee response refers to dispositions in LAR Attachment V, Table V-i, which,while indicating how the licensee addressed the related findings and observations (F&Os), do not specifically explain why failing to meet CC-Il is acceptable fortransition under NFPA 805. Provide Table V-2 which explains the rationale foracceptability of less than CC-Il satisfaction for all 22 SRs.RESPONSE:

During the peer review for the Farley fire PRA, peer reviewers identified 22 SRsmeeting a Capability Category less than CC-Il. Among them, only the following 3SRs had not been closed before the submittal of the LAR.* FSS-Ci* FSS-C2* FSS-E3These 3 SRs are met at CC-I level. As explained in Table V-2 attached to thisresponse, meeting these 3 SRs at CC-I level was determined to be acceptable forthis application.

For the remaining 19 out of 22 SRs, the technical issues in the Facts andObservations (F&Os) had been resolved after the peer review and beforesubmittal of the LAR in order to make these SRs met at the CC-Il level.The details of the closure of these 19 SRs are presented in Attachment V, TableV-1 revised in Enclosure 5 of the submittal of supplemental information (NL-12-2566 dated December 20, 2012) which was requested by the NRC for theacceptance review.Note that PRA RAI 29 RAI dated July 8, 2013 (ADAMS Accession No.ML13176A093),

included SRs FSS-D7 and FSS-H5 as being categorized lessthan CC-Il. However, these two F&Os have been closed and are considered tobe met at CC-Il as discussed below.There are 3 topics on SR FSS-D7; 1) the credited suppression system to beinstalled and maintained in accordance with applicable codes and standards, 2)floor non-suppression probability of 0, and 3) applicability of control room lambdato the MCR equipment rooms. The first two topics are discussed in the revisedTable V-I. Regarding the applicability of control room lambda to the MCRequipment rooms, Control Room lambda (0.33) was not used for MCR Equipment Rooms. Lambdas of electrical cabinet fire (0.102) and transient fires (0.126) areused for MCR equipment rooms.The F&O topic of SR FSS-H5 is the documentation of the uncertainty related tothe fire modeling for the fire scenarios.

Note that there is an F&O issued for SRFSS-E3 which is directly related to FSS-H5, but inconsistent with FSS-H5. Theindicated resolution for FSS-E3 states in part that the analysis documentation E4 -5 Enclosure 4Response to Probabilistic Risk Assessment RAIsshould be enhanced to note that methods for developing the statistical representation of the uncertainty intervals and mean values currently do not exist.However, F&O FSS-H5 then asks to undertake evaluations to addressuncertainty.

Since the documentation (Table D-1 of the Farley Fire PRASummary report, PRA-BC-F-1 1-017) has been updated to include discussions related to the uncertainty for fire modeling in response to SR FSS-E3, it isbelieved that SR FSS-H5 is closed.E4 -6 Enclosure 4Response to Probabilistic Risk Assessment RAIsFarley PRA RAI 33.a.01In a letter dated October 30, 2013 (ADAMS Accession No. ML1 3305A1 05), thelicensee responded to PRA RAI 33.a and referenced PRA RAI 16.a. However,neither of the responses to PRA RAI 16.a or PRA RAI 33 discussed the timing fordetection and manual suppression prior to fire spread to adjacent cabinets.

Furthermore, the response to PRA RAI 33.a indicates that all MCB panels arephysically open to one another.

Discuss the basis for assuming rapid enoughdetection and manual suppression prior to fire spread into the adjacent cabinet.(See also PRA RAI 16.a.01.)

RESPONSE

The response to this RAI will be provided by May 23, 2014.E4 -7 Enclosure 4Response to Probabilistic Risk Assessment RAIsFarley PRA RAI 33.c.01In a letter dated November 12, 2013 (ADAMS Accession No. ML13318A027),

thelicensee responded to PRA RAI 33.c and indicated an intent to revise its MCRabandonment calculation as follows:"The CCDP for the abandonment scenario is based on failure of allactions in the control room. [A] conservative basis was used fordetermining the abandonment CCDP based on the calculated CCDPassociated with panel damage and failure of the MCR actions.

The intentof this criteria is to ensure that the abandonment CCDP is an appropriate bounding value given that, shutting down the plant from outside thecontrol room has an inherently higher risk associated with it."These criteria are presented as (1) using conditional core damage probability (CCDP) = 0.1 if FRANC calculates a CCDP < 0.001, (2) using CCDP = 0.2 ifFRANC calculates a CCDP between 0.001 and 0.1, and (3) using 1.0 if FRANCcalculates a CCDP > 0.1. These FRANC-calculated CCDPs are based on bothMCB panel damage and failure of human actions in the MCR. Clarify how thesehuman actions were quantified, including any detrimental effects (increased failure probabilities) due to fire effects in the MCR. If screening or other boundingvalues were used, specify their bases, e.g., screening/scoping approach fromNUREG-1 921, "Fire Human Reliability Analysis Guidelines" (or equivalent).

RESPONSE

The response to this RAI will be provided by May 23, 2014.E4 -8 Enclosure 4Response to Probabilistic Risk Assessment RAIsFarley PRA RAI 35Section 2.4.3.3 of the NFPA 805 standard incorporated by reference into 10 CFR50.48(c) states that the PSA approach,

methods, and data shall be acceptable tothe AHJ, which is the NRC. Regulatory Guide (RG) 1.205, "Risk-Informed, Performance-Based Fire Protection for Existing Light-Water Nuclear PowerPlants,"

identifies NUREG/CR-6850 as documenting a methodology forconducting a Fire PRA (FPRA) and endorses, with exceptions and clarifications, NEI 04-02, "Guidance for Implementing a Risk-Informed, Performance-Based Fire Protection Program Under 10 CFR 50.48(c),"

Rev. 2, as providing methodsacceptable to the staff for adopting a fire protection program consistent withNFPA 805. RG 1.200, "An Approach for Determining the Technical Adequacy ofProbabilistic Risk Assessment Results for Risk-Informed Activities,"

describes apeer review process utilizing an associated ASME/ANS standard (currently ASME/ANS-RA-Sa-2009, "Addenda to ASME/ANS RA-S-2008, Standard forLeveil/Large Early Release Frequency Probabilistic Risk Assessment forNuclear Power Plant Applications")

as one acceptable approach for determining the technical adequacy of the PRA once acceptable consensus approaches ormodels have been established.

In a letter dated July 12, 2006 to NEI (ADAMSAccession No. ML061660105),

the NRC established the ongoing FAQ processwhere official agency positions regarding acceptable methods can bedocumented until they can be included in revisions to RG 1.205 or NEI 04-02.Section 2.4.4.1 of NFPA 805 states that the change in public health risk arisingfrom transition from the current fire protection program to an NFPA 805 basedprogram, and all future plant changes to the program, shall be acceptable to theAHJ, which is the NRC. RG 1.174, "An Approach for Using Probabilistic RiskAssessment in Risk-Informed Decisions on Plant-Specific Changes to theLicensing Basis," provides quantitative guidelines on CDF and LERF andidentifies acceptable changes to these frequencies that result from proposedchanges to the plant's licensing basis and describes a general framework todetermine the acceptability of risk-informed changes.As stated on page B-1 of Appendix B of PRA-BC-F-1 1-004, "Fire PRA LogicModel," the new Westinghouse Shutdown Shield (SDS) was installed in fall 2010.The internal events PRA (IEPRA),

upon which the FPRA is based, takes credit forthe SDS (failure rate of 0.0271/demand),

limiting the leakage rate to 2 gpm wherethe faces of the SDS seal components remain in contact.

The assumed leakagerate is increased to 19 gpm if the SDS actuates but the pump shaft continues torotate if not tripped in a timely manner. Finally, if the SDS does not actuate at all,"existing" (Westinghouse Owners Group (WOG) 2000 or Rhodes Model) sealmodel leakage rates are applied.

Given the July 26, 2013, 10 CFR Part 21notification by Westinghouse concerning defects with the SDS performance, provide a sensitivity evaluation that removes all credit for the SDS package,including both probability and consequences as appropriate.

Provide revisedestimates of CDF, LERF, A CDF and A LERF, including non-fire hazards for CDFand LERF, as a result of removal of this credit. Should this result in any changesto conclusions regarding the transition satisfying RG 1.174 risk/A risk guidelines, address any changes that will be made to accommodate this.E4 -9 Enclosure 4Response to Probabilistic Risk Assessment RAIsWhen performing this analysis, include the composite effect from all previous re-evaluations, including any synergistic

effects, specifically including the following:

a. From the LAR and the December 20, 2012 LAR Supplement, sensitivities related to the electrical cabinet fire severity method(Section V.2.1) and use of control power transformer (CPT)(Section V.2.3; also response to PRA RAI 08.a).b. From the RAI Responses dated September 16, 2013 (ADAMSAccession No. ML14038A019):

i. PRA RAI 01 .a -Removal of credit for VEWFDS in theMCR (also PRA RAI 01.01)ii. PRA RAI 15.a -Revised seismic CDF based on 2008USGS dataiii. PRA RAI 28.k -Validity of current Ignition Bin 15 firefrequencies

c. From the RAI Responses dated November 12, 2013 (ADAMSAccession No. ML13318A027):

i. PRA RAI 07.e -Use of 0.1 CCDP for MCR Abandonment ii. PRA RAI 17.d-Turbine Building Collapseiii. PRA RAI 33.c -Revised MCR Abandonment analysis (alsoRAI PRA 33.c.01)RESPONSE:

The response to this RAI will be provided by May 23, 2014.E4 -10 Joseph M. Farley Nuclear PlantResponse to Request for Additional Information Regarding License Amendment Request for Transition to 10 CFR 50.48(c)NFPA 805 Performance Based Standard for Fire Protection for Light WaterReactor Generating PlantsAttachment MRevision to License Conditions Farley -NFPA 805 LAR Markup -Attachment M -p. 4 of Attachment to Facility Operating Licensealternative to the Chapter 3 element is functionally equivalent or adequate forthe hazard. The licensee may use an engineering evaluation to demonstrate that a change to an NFPA 805, Chapter 3, element is functionally equivalent tothe corresponding technical requirement.

A qualified fire protection engineershall perform the engineering evaluation and conclude that the change hasnot affected the component, system, procedure, or physical arrangement functionality using a relevant technical requirement or standard.

The licensee may use an engineering evaluation to demonstrate that changesto certain NFPA 805, Chapter 3, elements are acceptable because thealternative is "adequate for the hazard."

Prior NRC review and approvalwould not be required for alternatives to four specific sections of NFPA 805,Chapter 3, for which an engineering evaluation demonstrates that thealternative to the Chapter 3 element is adequate for the hazard. A qualified fire protection engineer shall perform the engineering evaluation andconclude that the change has not affected the component, system,procedure, or physical arrangement functionality using a relevant technical requirement or standard.

The four specific sections of NFPA 805, Chapter 3,are: I 'nsert Text: This License Condition does not apply toany demonstration of equivalency under Section 1.7 of* Fire Alarm and Detection Systems (Section 3.8); NFPA 805.* Automatic and Manual Water-Based Fire Suppression Systems (Section3.9);* Gaseous Fire Suppression Systems (Section 3.10); and,P Passive Fire Protection Features (Section 3.11).(2) Fire Protection Program Changes that Have No More than Minimal RiskImpactPrior NRC review and approval are not required for changes to the licensee's fire protection program that have been demonstrated to have no more than aminimal risk impact. The licensee may use its screening process asapproved in NRC safety evaluation report dated to determine that certain fire protection program changes meet the minimal criterion.

Thelicensee shall ensure that fire protection defense-in-depth and safety marginsare maintained when changes are made to the fire protection program.Transition License Conditions (1) Before achieving full compliance with 10 CFR 50.48(c),

as specified by (2)below, risk-informed changes to the licensee's fire protection program maynot be made without prior NRC review and approval unless the changehas been demonstrated to have no more than a minimal risk impact, asdescribed in (2) above.(2) The licensee shall implement the following modifications to its facility tocomplete transition to full compliance with 10 CFR 50.48(c) bySee plant specific list of modifications identified in Attachment S.(3) The licensee shall maintain appropriate compensatory measures in placeuntil completion of the modifications delineated above.

Joseph M. Farley Nuclear PlantResponse to Request for Additional Information Regarding License Amendment Request for Transition to 10 CFR 50.48(c)NFPA 805 Performance Based Standard for Fire Protection for Light WaterReactor Generating PlantsAttachment VRevision to Fire PRA Quality Southern Nuclear Operating CompanyAttachment V -Fire PRA QualityTable V-2 Fire PRA- Category I Summary1SR Capability F&O # and Finding/Observation StatusCategoryFSS-C1 CC-I [FSS-C1-011 Two-point fire intensity model that encompass low likelihood, butpotentially risk contributing, fire events were not used in allcases. Fire scenarios were done with ignition sourcescharacterized with one fire intensity.

To reach Capability Category II, use a two-point intensity modelfor all ignition sources.Utility Comment:

The development of fire scenarios for theFarley Fire PRA did not identify any instances where furtheranalysis resolution would be gained by the treatment as inferredby the requirements for CC II and CC Ill. The implications ofretaining the CC I treatment in lieu of refining as described forCC II or CC III is potentially a higher calculated CDFcontribution.

The CC I treatment inherently will not result inunder-estimation of fire risk. As such, the current treatment isconservative.

Provided this treatment does not result in maskingof risk increases in future applications, further refinements arenot considered necessary.

The development of fire scenarios for the Farley Fire PRA did notidentify any instances where further analysis resolution would begained by the treatment as inferred by the requirements for CC IIand CC I1l. The implications of retaining the CC I treatment in lieuof refining as described for CC II or CC Ill is potentially a highercalculated CDF contribution.

The CC I treatment inherently willnot result in under-estimation of fire risk. As such, the currenttreatment is conservative.

Provided this treatment does not resultin masking of risk increases in future applications, furtherrefinements are not considered necessary.

Response:

The SR stipulates that a two-point model is requiredfor CC-Il. As you stated in your comment, Farley feels that theone-point model is conservative and justified.

This would beviewed as the proposed resolution, but the F&O stands.FSS-C2 CC-1[FSS-C2-01]

Ignition source intensity were characterized such that fire isinitiated at full peak intensity and ignition sources that aresignificant contributors to fire risk were not characterized using arealistic time-dependent fire growth profile.

Generic methodsfrom the Hughes Associates Generic Fire Modeling Treatments were used to characterize ignition source intensity.

Thesegeneric methods did not incorporate fire growth curves.The only readily available reference for a time dependent growthrate that could be considered in the analysis is 12 minutes asrecommended in NUREG/CR-6850.

The treatment would involvea t2 growth rate. If a particular source/target interaction has aspacing where the target is at the critical damage spacingthreshold, such a treatment may provide some benefit assuccessful suppression with that time period would prevent targetdamage. However, if the target is located well within theRev 01 Page V-31Rev 01Page V-31 Southern Nuclear Operating CompanyAttachment V -Fire PRA QualityTable V-2 Fire PRA- Category I Summary1SR Capability F&O # and Finding/Observation StatusCategoryCharacterize ignition sources that are significant contributors tofire risk using a realistic time-dependent fire growth profile.Utility Comment:

The only readily available reference for atime dependent growth rate that could be considered in theanalysis is 12 minutes as recommended in NUREG/CR-6850.

The treatment would involve a t2 growth rate. If a particular source/target interaction has a spacing where the target is at thecritical damage spacing threshold, such a treatment may providesome benefit as successful suppression with that time periodwould prevent target damage. However, if the target is locatedwell within the calculated damage distance, the corresponding time to reaching the damage threshold is very short andeffectively precludes any meaningful credit for suppression.

Inthe case of the Farley Fire PRA, the majority of the targetspacing for the dominant risk contributors is such that nomeaningful credit for suppression is available.

In other dominantrisk contributors, the scenario involves high energy arcing fault(HEAF) events were no growth time is applicable.

Theimplications of retaining the CC I treatment in lieu of refining asdescribed for CC I1/111 is potentially a slightly higher calculated CDF contribution.

The CC I treatment inherently will not result inunder-estimation of fire risk. As such, the current treatment isconservative.

Provided this treatment does not result in maskingof risk increases in future applications, further refinements arenot considered necessary.

calculated damage distance, the corresponding time to reachingthe damage threshold is very short and effectively precludes anymeaningful credit for suppression.

In the case of the Farley FirePRA, the majority of the target spacing for the dominant riskcontributors is such that no meaningful credit for suppression isavailable.

In other dominant risk contributors, the scenarioinvolves high energy arcing fault (HEAF) events were no growthtime is applicable.

The implications of retaining the CC I treatment in lieu of refining as described for CC Il/111 is potentially a slightlyhigher calculated CDF contribution.

The CC I treatment inherently will not result in under-estimation of fire risk. As such, the currenttreatment is conservative.

Provided this treatment does not resultin masking of risk increases in future applications, furtherrefinements are not considered necessary.

Response:

The Farley modeling was found to be consistent with CC-I but did not meet the requirements of CC-Il. Thecomment provides the basis for stating that the existingtreatment is adequate.

It does not provide evidence that a time-dependent heat release rate model was used.FSS-E3 CC-I [FSS-E3-01]

Supporting requirement E3 asks to provide a mean value of, and The documentation has been updated to include discussions Rev 01Page V-32 Southern Nuclear Operating CompanyAttachment V -Fire PRA QualityTable V-2 Fire PRA- Category I Summary1SR Capability F&O # and Finding/Observation StatusCategorystatistical representation of, the uncertainty intervals for the related to the uncertainty for fire modeling.

See Table D-1 of theparameters used for fire modeling the fire scenarios.

Farley Farley Fire PRA Summary report, PRA-BC-F-1 1-017.performed fire size and heat release rate selection inaccordance with NUREG/CR-6850 and/or applicable FAQs. The associated SR was dispositioned as CC I which is judged toHowever, the methods for developing the statistical be sufficient given the two concerns noted.representation of the uncertainty intervals and mean valuescurrently do not exist. However, this is not reported in thedocumentation.

In the documentation, explain that it is understood that methodsfor developing the statistical representation of the uncertainty intervals and mean values currently do not exist.Utility Comment:

This specific F&O was issued against atechnical element and the indicated resolution involves adocumentation clarification.

This documentation clarification willbe implemented.

1 _ All Fire PRAs SRs characterized as Capability Category I were identified as Findings in the Peer Review.Rev 01 Page V-33Rev 01Page V-33